Review on Soft Mobility Infrastructure Design Codes

Abstract

Soft mobility is gaining popularity in urban spaces due to its various benefits in terms of carbon footprint, air quality, congestion mitigation, and public health. Soft mobility infrastructure mainly includes urban road adjustments to accommodate pedestrian and bicycle flows. Relevant design codes are being developed worldwide, and important investments are being made in soft mobility. This paper provides a review and comparative analysis of 17 design codes and regulations from different countries and regions across the world. Furthermore, the German road design code for motorized traffic is used as a reference to assess the level of detail and eventual gaps in the soft mobility infrastructure design codes. Results indicate that, in contrast to road codes, soft mobility infrastructure codes vary significantly from country to country. Most importantly, the limit and recommended values of geometric parameters are fewer in number and less documented compared to road design parameters. Evidence-based recommendations are needed to enhance the design, construction, operation, maintenance, and safe management of soft mobility infrastructure.

1. Introduction

As urbanization and transportation modes continue to change and develop, many derivative problems are raised [1]. A US study shows that from 1982 to 2011, the total traffic congestion time in the top 101 major cities in the United States increased by 4.8 billion hours [2]. At the same time, traffic congestion may further aggravate environmental pollution, such as greenhouse gas emissions [3]. However, environmental pollution can also cause people’s health problems [4]. In this context, infrastructure for soft mobility has become a key focus in the field of urban planning and sustainable transportation. Soft mobility encompasses non-motorized transportation, including walking, cycling, roller skating, and skateboarding. It can also be viewed as “zero-impact” mobility. While terms to describe this mode of movement are not standardized yet, phrases like “soft mobility”, and “slow traffic” are used interchangeably, primarily emphasizing pedestrians and cyclists as alternatives to cars [5].

Walking and cycling complement and connect with each other. A walk will inevitably follow before and after the ride. Cycling and walking therefore share similar advantages and contributions to society. Taking cycling as an example, it has many advantages for both cyclists and society: it is a low-cost, low-polluting, health-improving way to travel [6]. Cycling is more space-efficient than car transport, environmentally friendly, and requires minimal investment (excluding professional sports). Operational costs are negligible as bicycles do not need fuel, and expenses like energy beverages for cyclists do not significantly impact on the overall cost. Additionally, cycling offers door-to-door transport, unlike public transportation [7]. Especially during COVID-19 pandemic outbreaks, cycling in an open-air environment can avoid crowds gathering while providing a good ventilation environment, which is conducive to the dissipation of the virus and protects people’s health [8].

The resilience of cycling infrastructure has emerged as a critical component in fostering sustainable urban mobility. Firstly, enhancing infrastructure design to prioritize road safety is imperative, particularly for vulnerable road users who disproportionately bear the risks of traffic accidents. For instance, segregated bicycle lanes with physical barriers (e.g., bollards or curbs) and optimized intersection signals can reduce conflicts with motorized traffic, thereby lowering injury rates. Secondly, economic sustainability must be embedded in lifecycle planning. Integrating quality standards into regulations ensures long-term maintenance efficiency and cost-effectiveness [9]. Prioritizing durability during the design phase can reduce long-term fiscal burdens while maintaining operational integrity. Notably, cycling infrastructure has proven to be a resilient alternative during crises. The COVID-19 pandemic highlighted its adaptability, with cities especially in South Korea and Japan observing a modal shift from public transit to bicycles [10,11]. Temporary infrastructure expansions, such as emergency bike lanes, not only accommodated increased demand but also alleviated pressure on overwhelmed transport systems. This dual role—serving daily commutes and emergency mobility—underscores cycling infrastructure’s capacity to enhance systemic resilience, ensuring continuity during disruptions while advancing climate-responsive urban agendas.

While human-driven transportation is one possibility, it does not encompass all soft mobility modes, necessitating other alternatives powered by pure energies. Electric-powered two-wheelers, like e-scooters and e-bikes, serve as such alternatives [12]. E-scooters have recently emerged as a promising sustainable mode for urban trips, and their popularity is rapidly growing across the world [13]. They may be used either for short door-to-door trips or for first- or last-mile connections to public transport stops and stations. They may be privately owned or part of large, shared fleets. Their extensive usage has raised various challenges regarding the rights and obligations of their riders and vehicle specifications and requirements. In addition, a lot of public debate is being held around safety concerns [14]. E-scooter riders are particularly vulnerable to motorized traffic when using mixed traffic lanes. When using soft mobility infrastructure (SMI), e-scooter riders put other road users, such as pedestrians, at risk. In both cases, transportation infrastructure (be it motorways or SMI) is certainly not designed for e-scooters, and the characteristics of those new vehicles are not properly considered. Their size and dynamic characteristics, such as speed and acceleration rates, are significantly different from those of all other road users.

Regulations, standards, and design specifications for SMI differ significantly across countries and regions. The objective of this paper is to provide an overview and comparative analysis of the current state of the practice. Comparison to road infrastructure design specifications is made to underline the differences and highlight important gaps. This paper has multiple utilities as it offers a comprehensive overview of codes and regulations for engineers and scientists, a reference to the best international practices for practitioners, funding priority guidance for decision-makers, and the identification of research gaps for researchers.

The remainder of this paper is organized as follows: Section 2 provides a cross-national and cross-regional review of SMI design codes, laws, and regulations. Section 3 gives the main conclusions and some recommendations by the review. Supplementary File S1 shows the summary of the regulations and design standards of the countries and regions been reviewed. Supplementary File S2 presents the relevant German road design specifications that are uses as a benchmark for SMI evaluation.

2. Codes for Soft Mobility Infrastructure

2.1. Introduction

In the review of SMI, data were collected from 17 different countries, regions, or cities around the world that have issued special laws, regulations, design specifications, and guidance manuals for bicycle and pedestrian facilities. Figure 1 and

Table 1. Document type of countries, regions, and cities reviewed and road length data.

Country	Area of Application		Code	Document Type	Document Reference	Total Road Network Length * (km)	Road Length per Capita * (m/person)	Road Length per km2 * (m/km2)
Kingdom of Denmark	Nationwide		DK	Handbook	[16]	74,558	12.92	1757.0
Republic of Austria	Nationwide		AT	Guidelines	[17,18]	137,039	15.50	1662.2
Hellenic Republic (Greece)	Nationwide		GR	Instruction	[19]	117,000	10.90	895.5
New Zealand	Nationwide		NZ	Standard and Guidelines	[20]	94,000	19.61	355.3
Republic of Singapore	Nationwide		SG	Design Guidelines	[21]	3500	0.62	4935.1
Japan	Nationwide		JP	Government Order	[22]	1,218,772	9.59	3343.8
Republic of South Africa	Nationwide		ZA	Guidelines	[23]	750,000	13.48	617.6
United Republic of Tanzania	Nationwide		TZ	Design Manual	[24]	87,581	1.80	98.9
Russian Federation	Nationwide		RU	Rules	[25]	1,283,387	9.01	78.4
People’s Republic of China	Nationwide		CN	Design Standard	[26]	4,960,600	3.58	531.9
	Local	Beijing	BJ	Design Standard	[27]	22,559 [28]	1.03 +	1374.7 +
	Local	Hong Kong	HK	Design Standard and Guidelines	[29]	2107	0.29	1963.7
	Local	Taiwan	TW	Design Guidelines	[30]	43,206	1.83	1339.3
United States of America	Local	Washington D.C.	WA	Design Manual	[31,32]	2293 [33]	3.22 +	12,954.8 +
	Local	New Jersey	NJ	Design Guidelines	[34,35]	62,683 [36]	6.60 +	2775.0 +
Republic of India	Local	Delhi	DL	Guidelines	[37]	33,198 [38]	1.98 +	22,360 +
United Arab Emirates	Local	Abu Dhabi	AD	Design Manual	[39]	26,997 [40]	7.10 +	401.0 +

* Except individual marks, road length data are cited from [41]. Numbers with ⁺ means they are calculated by population and area.

illustrate the countries, regions, and cities that cover both developed and developing countries and include the European Union, ASEAN, the Commonwealth, and various inter-state alliances. They refer to countries with different population densities, landscapes, political systems, and policy-making processes. The naming of countries follows the two-character codes provided by ISO-3166-1 ALPHA-2 [15]. Regarding local provinces, states, and cities, two-letter codes were given by the authors. This coding is used in all tables and figures as well as text that follow. The soft mobility infrastructure design standards and codes of the 17 countries and regions analyzed are summarized in a table in the Supplementary File S2. Also, the data of the total length of road network, as well as the road length per capita and per km², are also shown in Table 1, in order to show the road construction status of these countries and regions.

In what follows, the design standards appearing in the codes are organized and comparatively assessed. The analysis includes the following elements that are presented in different sections and subsections: (i) classifications, (ii) design features (layout, vertical and horizontal geometrics, speeds, and stopping sight distance), (iii) roadside facilities (traffic guidance facilities, lightning, weather protection, and traffic counter), and (iv) supporting facilities (parking, end-of-trip facilities).

2.2. Classifications

The right-of-way (ROW) is the right of a vehicle or pedestrian to proceed uninterruptedly in a lawful manner in the direction [42]. Five main categories in terms of ROW classification were identified in the reviewed codes: (i) shared bicycle lanes with motor vehicles, (ii) advisory bicycle lane in the motor vehicle lane, (iii) curb-divided bicycle lane, (iv) dedicated bicycle lane buffered by road marking, and a (v) dedicated bicycle lane buffered by green belt. They may all be one- or two-way and they may follow car traffic direction or operate contra-flow. In actual operations, some countries and regions use cross-combination types based on their actual traffic needs. In Washington, D.C., for example, two-way buffer-divided bicycle lanes on only one side of the road and contra-flow bike lanes on one-way motor roads are applied.

2.3. Design Features

2.3.1. General Layout

The laws, regulations, and design standards of many countries and regions around the world stipulate the minimal width of bicycle lanes as it plays a vital role in the capacity and safety of SMIs [43]. Most countries and cities (NZ, GR, DK, BJ) have detailed regulations on lane width. A typical cross-section plan of bicycle and pedestrian SMI widely used appears in Figure 2.

Where, a—Curb or buffer that divides the bicycle and motor vehicles, b—Width of bicycle lane or path, b₀—Physical width of a bicycle, b₁—Lateral clearance for the bicycle lane or path, b₂—Spacing for bicycles, h_b—Height clearance for bicycles, p—Width of pedestrian passage or path, p₀—Physical width of a pedestrian or a wheelchair, p₁—Lateral clearance for the pedestrian passage or path with bicycle lane or path (“street furniture and greening zone”), p₂—Spacing for pedestrians, p₃—Lateral clearance for the pedestrian passage or path with buildings (“frontage”) and h_p—Height clearance for pedestrians.

The existence and eventual value of an upper limit for SMI lane width is not based on evidence but may be directly decided by relevant authorities. The minimal width value comes from the estimation of the dynamic width of the moving object. Specifically for one-way bicycle lanes, the minimum width requirement varies from 0.75 m (NZ) to 3.5 m (GR), with the most common value (the dominant value that appears in more than half of the analyzed codes) being 2.0 m. For two-way bicycle lanes, the minimum width requirement varies from 2.0 m (DK) to 4.7 m (NZ), with the most common value being 3.0 m. As for pedestrian paths, the minimum width requirement varies from 0.9 m (AT) to 4.5 m (HK), with the most common value being 3.0 m.

Meanwhile, there are also recommended width given in the codes. For one-way bicycle lanes, this width varies from 1.5 m (JP) to 4.5 m (BJ), with the most common value being 2.5 m. For two-way bicycle lanes, it varies from 2.5 m (AD) to 4.8 m (NZ), with the most common value being 3.5 m. As for pedestrian paths, it varies from 1.5 m (TZ) to 5.0 m (BJ) with the most common value being 3.5 m.

However, some countries and regions do have specific formulas to determine lane width, such as the CN formulas shown as Formulas (1) and (2) [26].

For pedestrian:

$$\begin{array}{c}{W}_p={\displaystyle \frac{N_w}{N_{w1}}}\times W_1\end{array}$$

(1)

For non-motor vehicles:

$$\begin{array}{c}{W}_b={\displaystyle \frac{N_b}{N_{b1}}}\times W_2+0.25\times 2\end{array}$$

(2)

where

• W_p—Width of the pedestrian lane (m);
• N_w—Peak-hour pedestrian flow on the pedestrian lane (ped/h);
• N_w1—Design capacity of a single pedestrian lane (ped/h);
• W₁—The width of a single pedestrian lane (m) which refers to the b₀ in Figure 2;
• W_b—Width of the non-motor vehicle (m);
• N_b—Peak-hour non-motor vehicle flow on non-motor vehicle lane (veh/h);
• N_b1—Design capacity of a single bicycle lane (veh/h);
• W₂—The width of a single bicycle lane (m) which refers to the p0 in Figure 2;
• 0.25—Lateral clearance of the non-motor vehicle lane (m).

Clearance height refers to the minimum vertical space required to ensure the safe passage of pedestrians and bicycles. Unlike motor vehicles, which are calculated by “fixed height + safety margin”, the height of pedestrians and cyclists has dynamic characteristics, so the determination of this indicator needs to consider both human characteristics and safety needs. Countries generally use empirical numerical methods to set it: first determine the basic parameters based on the standard height of the adult population in the country and then add the dynamic safety margin. The final clearance height value is usually significantly lower than the longitudinal safety space standard for motor vehicles (4.7 m, as shown in Figure S1 in the Supplementary Materials). This differentiated setting logic stems from the essential difference between pedestrians and vehicles in terms of movement trajectory and space occupancy. For specific numerical standards, please refer to the comparative analysis of data from major countries/regions around the world in

Table 2. Requirements for bicycle lanes, pedestrian paths, and other road components across the world (units in [m]).

Code	Bicycle Lane													Pedestrian Path
	One-Way						Two-Way or Non-Directional *						hb
	b0		b1		b2		b0		b1		b2			p0		p1		p2		p3		hp
	Min	Max	Min	Max	Min	Max	Min	Max	Min	Max	Min	Max	Min	Min	Max	Min	Max	Min	Max	Min	Max	Min
AD	-	-	-	-	-	-	1.5	2.5	-	-	-	-	-	1.8	4.0	1.0	3.5	-	-	0.3	1.5	-
AT	1.0	2.6	0.5	1.0	-	-	2.0	4.0	0.5	1.0	-	-	2.25	0.9	2.0	0.3	4.25	-	-	0.3	4.25	2.2
BJ	3.5	-	-	-	-	-	3.0	4.5	-	-	-	-	-	1.5	5.0	-	-	-	-	-	-	-
CN	3.5	4.5	-	-	0.25	-	2.5	4.5	-	-	0.25	-	2.5	2.0	5.0	-	-	-	-	-	-	2.5
DK	1.5	-	-	-	-	-	2.0	3.0	-	-	0.5	-	-	2.5	-	-	-	-	-	-	-	-
DL	-	-	-	-	-	-	1.5	2.5	-	-	-	-	-	1.8	-	1.8	-	-	-	0.5	1.0	2.4
GR	1.5	4.0	0.1	1.0	-	-	2.5	4.0	0.1	1.0	-	-	-	-	-	-	-	-	-	-	-	-
HK	2.0	2.8	-	-	-	-	3.5	4.0	-	-	-	-	-	2.0	4.5	1.5	4.0	-	-	0.5	1.0	-
JP	-	-	-	-	-	-	1.0	4.0	-	-	-	-	-	2.0	3.5	-	-	-	-	-	-	-
NJ	-	-	1.4	1.5	-	-	1.2	2.4	1.4	1.5	-	-	-	0.9	3.0	0.6	2.4	-	-	0.4	0.8	2.03
NZ	1.6	3.0	-	-	-	-	1.6	4.0	-	-	-	-	-	1.5	3.0	0	2.5	-	-	0	1.0	2.0
RU	-	-	-	-	-	-	0.75	2.0	-	-	-	-	-	2.0	4.0	-	-	-	-	-	-	-
SG	-	-	-	-	-	-	1.5	4.0	-	-	-	-	-	1.5	1.8	2.0	-	-	-	2.4	5.0	-
TW	1.5	2.0	-	-	-	-	2.0	3.0	-	-	-	-	2.5	0.9	2.5	-	-	-	-	-	-	2.1
TZ	-	-	-	-	-	-	1.5	2.5	-	-	-	-	1.9	1.0	3.5	-	-	-	-	-	-	-
WA	-	-	-	-	-	-	1.2	3.6	-	-	-	-	-	1.2	4.8	1.2	3.0	-	-	1.8	4.5	-
ZA	-	-	0.6	-	-	-	1.2	1.8	0.6	-	-	-	-	1.2	3.5	-	-	-	-	-	-	2.1

* Non-directional means a bicycle path for both directions without assignment and bicycle lanes on the side of motor vehicle lanes.

.

However, research also shows that the “eye height” of the driver significantly affects the sight and estimation of vertical slop length and horizontal vertical radii of the driver. Body height is an important factor affecting the “eye height” of the driver [44,45,46]. Compared to the data from a 2017 article that studied the changing trends in human height around the world between 1896 and 1996 [47], the previous results show that there is no obvious trend similarity between the height clearance of pedestrian paths required by various countries and the average height of their citizens in 1996 (

Table 3. Average citizen height in 1996 and pedestrian height clearance required in DL, ZA, CN, TW, HJ, AT, and NZ.

Country Code	Average Citizen Height in 1996 (m)	Height Clearance Requirement for Motor Vehicles (m)
AT	1.775	2.20
CN	1.720	2.50
DL	1.655	2.40
NJ	1.770	2.03
NZ	1.780	2.00
TW	1.750	2.10
ZA	1.665	2.10

). A more obvious example is that NZ has the highest average height of its citizens but the lowest requirement for pedestrian path height clearance. Meanwhile, CN has the third lowest average height of its citizens but the highest requirement for pedestrian path height clearance.

Three determination methods were identified from the reviewed countries and regions:

(i)

Based on the estimated traffic volume of the SMI itself (e.g., NZ, DK, AT, GR, HK, TW, JP, SG, DL, and TZ). Indicatively, Figure 3 gives the width requirement of footpaths in New Zealand as a function of estimated pedestrian flows.

(ii)

Based on the speed limit and/or volume of motor vehicle roads adjacent to the SMI (i.e., NJ, NZ, CN, BJ, and AD), with larger widths being required for upper-class roads. Indicatively,

Table 4. Shoulder seal width for state highway cycling network of New Zealand [20].

Speed Limit of Adjacent Traffic Lane	50 km/h	70 km/h	100 km/h
Minimum adjacent traffic lane width	3.0 m	3.3 m	3.5 m
1–1000 AADT	0.0 m	0.0 m	0.0 m
1000–2000 AADT	0.75 m	0.75 m	0.75 m
2000–5000 AADT	1.0 m	1.0 m	1.0 m
5000–8000 AADT	1.2 m	1.5 m	1.5 m
8000–18,000 AADT	1.5 m	1.7 m	2.0 m
18,000+ AADT	2.0 m	2.0 m	2.2 m

provides cycle lane widths according to Average Annual Daily Traffic (AADT) and speed limits in NZ.

(iii)

Based on where the SMI is located (i.e., AD, WA, and ZA), with larger widths required for urban commercial areas.

Table 5. Width for pedestrian and bicycle infrastructure in Abu Dhabi, UAE (units in [m]) [39].

Road Location	Road Type	Pedestrian Passage				Bicycle Lane/Track
		Pedestrian Through		Cycle Track at Pedestrian Realm		Bicycle Lane on Frontage Lane		Bicycle Lane on Traveled Way
		Min	Max	Min	Max	Min	Max	Min	Max
City	Boulevard	2.8	4.0	1.5	2.5	2.5	3.0	-	-
	Avenue	2.4	4.0	1.5	2.5	2.5	3.0	1.5	2.5
	Street	2.4	3.0	1.5	2.5	-	-	1.5	2.5
	Access Lane	1.8	2.5	-	-	-	-	-	-
Town	Boulevard	2.4	3.5	1.5	2.5	2.5	3.0	-	-
	Avenue	2.0	3.0	1.5	2.5	2.5	3.0	1.5	2.5
	Street	2.0	2.4	1.5	2.5	-	-	1.5	2.5
	Access Lane	1.8	2.5	-	-	-	-	-	-
Commercial Area	Boulevard	2.4	3.0	1.5	2.5	2.5	3.0	-	-
	Avenue	2.0	3.0	1.5	2.5	2.5	3.0	1.5	2.5
	Street	2.0	2.4	1.5	2.5	-	-	1.5	2.5
	Access Lane	1.8	2.5	-	-	-	-	-	-
Residential Area	Boulevard	1.8	2.8	1.5	2.5	2.5	3.0	-	-
	Avenue	1.8	2.0	1.5	2.5	2.5	3.0	1.5	2.5
	Street	1.8	2.0	1.5	2.5	-	-	1.5	2.5
	Access Lane	1.8	2.5	-	-	-	-	-	-
Industrial Area	Boulevard	2.0	3.6	1.5	2.5	3.7	4.0	-	-
	Avenue	2.0	3.4	1.5	2.5	3.7	4.0	1.5	2.5
	Street	2.0	3.0	1.5	2.5	-	-	1.5	2.5
	Access Lane	1.8	2.5	-	-	-	-	-	-
No Active Frontage	Boulevard	1.8	3.5	1.5	2.5	2.5	3.0	-	-
	Avenue	1.8	3.0	1.5	2.5	2.5	3.0	1.5	2.5
	Street	1.8	2.4	1.5	2.5	-	-	1.5	2.5
	Access Lane	1.8	2.5	-	-	-	-	-	-

is a relevant example for both cycling and pedestrian lanes from Abu-Dhabi.

Different from the design for motor vehicles, cross-slopes are generally designed in only one direction for SMI, towards the side of the motor vehicle lane, as shown in Figure 4. The cross-slope is only required in DK, ZA, NJ, WA, and TW for bicycle lanes and sidewalks. As a summary in

Table 6. Cross-slope requirements for SMI across the world.

Code	Target	Required Cross-Slope (%)
		Maximum	Minimum	Requirement
AD	-	-	-	-
AT	-	-	-	-
BJ	-	-	-	-
CN	-	-	-	-
DK	Bicycle	-	-	2.5
DL	-	-	-	-
GR	-	-	-	-
HK	-	-	-	-
JP	-	-	-	-
NJ	-	-	-	-
NZ	Bicycle	3	2.5	-
	Pedestrian	-	-	2
RU	-	-	-	-
SG	-	-	-	-
TW	Bicycle	2	0.5	-
	Pedestrian	5	0.5	-
TZ	-	-	-	-
WA	Pedestrian	2	-	-
ZA	Pedestrian	-	-	2

, in NZ and TW, cross-slopes of 0.5–3% are specified for pedestrian and bicycle paths to meet drainage needs. It can be seen in Table 6 that many countries have no specific requirement regarding the cross-slope of SMI, contrarily to road design codes. This is an important omission that may lead to infrastructure degradation due to poor rainwater management and, also, reduced rider comfort and safety levels.

Most countries and regions only give out data on lane width directly, without any basis or method to calculate it except CN. Meanwhile, the ways to determine the lane width across the world for SMI are different from the ways for motor vehicle lanes. Only parameters (i.e., volume, speed limit, and geometry) about the motor vehicle lanes are considered when determining the lane width, while when determining the lane width of SMI, only other participants are considered instead of the parameters of the SMI.

2.3.2. Vertical Geometrics

Vertical slope has a strong impact on the comfort of pedestrians and non-motor vehicle riders when using the infrastructure. Of course, the slope is generally affected by terrain conditions. Areas with more undulating terrain generally have looser slope regulations, while plain areas are more stringent. For example, in Chongqing, a well-known mountainous municipality in mainland China, certain provisions requiring slopes in the Chinese road design standards that are generally applicable nationwide can be exempted from application there. SG (4%), GR (5%), HK (10%), and NJ (3%) only have maximum vertical slope regulations for bicycle lanes or paths. ZA (5%), NZ (3%), and AD (5%) only have maximum vertical slope regulations for pedestrian passages or paths. Meanwhile, WA (5% and 5%) and TW (8% and 12%) have vertical slope regulations for both bicycle and pedestrian lanes or paths. The requirements are shown in

Table 7. Maximum slope gradient required across the world.

Code	Target	Vertical Slope Requirement (%)
		Acceptable	Maximum	Treatment When Greater than the Maximum
AD	Pedestrian	-	5	Treat as a ramp
AT	-	-	-	-
BJ	-	-	-	-
CN	-	-	-	-
DK	-	-	-	-
DL	-	-	-	-
GR	Bicycle	3	5	-
HK	Bicycle	3	10	-
JP	-	-		-
NJ	Bicycle	-	3	Additional space
NZ	Pedestrian	-	3	3–5, provide rest platforms 5–8, provide wayfinding signage >8, treat as a ramp
RU	-	-	-	-
SG	Bicycle	-	4	-
TW	Bicycle	5	8	-
	Pedestrian	5	12	-
TZ	-	-	-	-
WA	Bicycle	-	5	Additional width
	Pedestrian	-	5	Provide rest platforms
ZA	Pedestrian	-	5	-

. The same as Table 6, Table 7 shows us that there are many countries and regions do not have requirements about vertical slope gradient. This may result in the SMI road in reality not being able to provide the most comfortable experience for SMI users. In severe cases, it may even lead to traffic accidents such as slipping.

When the slope reaches a certain gradient, most countries and regions provide either “rest platforms” (NJ) or increase the lane or path width (WA) so that SMI users can take a break to recover after a long climb or wider space for climbing the hill at a low speed. As shown in Figure 5, a “rest platform” is a large horizontal open space, generally of the same width as the driveway, and the length meets basic needs.

However, especially in Taiwan, a maximal slope of 3% is only allowed for lengths of over 500 m. When the slope reaches 8%, the maximum slope length is not allowed to exceed 35 m. Otherwise, rest platforms need to be provided. In some countries, in areas where bicycle traffic and demand are very large, specific devices that use external forces to transport cyclists are implemented. As shown in Figure 6, the world’s first “bicycle elevator” was built in Trondheim, Norway, to help cyclists climb hills. This bicycle elevator is called a Trampe (Skirail, Trondheim, Norway), and it helps you get to the top of a slope easily without riding yourself. In 2012, Trampe was dismantled and replaced by a more industrial version, CycloCable (Skirail, Trondheim, Norway). The structure of CycloCable is very simple: it is a lifting channel embedded on the roadside. When going uphill, a cyclist puts one foot on the pedal provided by the channel, and it will push them and their bicycle uphill at a speed of 2 m/s [48].

The vertical slope design of SMI is different from motor vehicle roads. The vertical slope design of motor vehicle roads mentioned in Supplementary File S2 depends on the road classification, which is related to the speed limit and location environment of the road. The vertical slope of SMI is only related to slope length and has different maximum slope restrictions for different uses (maximum slope for pedestrians is lower than bicycles). Furthermore, due to differences in driving force and mechanical structure, the maximum gradient limit for pedestrians is much looser than that for bicycles.

2.3.3. Horizontal Geometrics

The radius of horizontal curvature influences the stability of bicycles and the driver’s speed choice [49]. Among the 17 cases viewed, only HK and TW give out specific limits on horizontal radii. HK regulations prefer minimum horizontal radii of 5 m, but in difficult conditions, radii of 2 m may be acceptable. TW specifies the minimum horizontal radii into the design speed of the bicycle paths, 3 m for 10 km/h, 15 m for 20 km/h, and 30 m for 30 km/h, which uses the same concept for motorway design.

The superelevation for bicycle lanes is largely similar to the need for superelevation for motor vehicle lanes, but there are some specific considerations. By setting appropriate superelevation for bicycle lanes, the stability of cyclists in curves can be improved and the possibility of cyclists falling or losing balance in curves can be reduced, thereby reducing the incidence of cycling accidents, as shown in Figure 7. At the same time, the superelevation of the bicycle lane can allow cyclists to pass curves faster, increasing the riding speed and the traffic capacity of the bicycle lane. However, among the 17 cases we have information on, only TW has relevant requirements for bicycle lanes superelevation with the optimal value of the superelevation of the bicycle lane being 2%, and the maximum cannot exceed 3%.

2.3.4. Speeds

As shown in Supplementary File S2, speed limit and design speed are the most important keys input in road design but rarely appear in SMI codes. This could be due to the lack of evidence as actual speeds remain largely unknown due to limited speed measuring equipment. Variables equivalent to V85 were not found in the reviewed codes. However, scientific evidence suggests that bicycle speeds do show variability, and they are affected by factors like cars. A recent study in China, for example, indicates that low bicycle volumes and larger number of male cyclists are related to higher average bicycle speeds [50]. Similarly, when the pedestrian density increases, the average walking speed of pedestrians will decrease until the density is too high and they cannot move [51].

On the contrary, bicycle speed limits and design limits are defined in BJ, TW, WA, and ZA based on safety criteria. Those limits vary from 20 km/h (BJ) to 50 km/h (ZA). There, BJ has a speed limit of 20 km/h for all kinds of cycle paths and lanes. Meanwhile TW carried out the design speed varies from 10 km/h to 30 km/h which will influence the horizontal radii of the path. WA only regulated the design speed of SMI of 18 mph (29 km/h). But for ZA, they made more specific for the SMI design speed. When the downhill grade is less than 3%, 30 km/h design speed is required, 40 km/h for 3–7%, and 50 km/h for more than 7%. It should be noted though that lower speed limits may apply to specific SMI if necessary. Finally, enforcement is not a common practice as speed radars capable of verifying violations at this speed range are not generally commercialized, traffic police do not make regular controls, and, in many cases, relevant fines are not defined in national road codes.

Speed limits for pedestrians were not observed. But, speed limits and special limitations only for e-scooters are required in the countries and regions we reviewed except BJ, NA, and TZ, as summarized in

Table 8. E-scooter usage requirements in countries and regions around the world.

Code	Minimum Age-Allowed	Helmet (Until Age)	Speed Limit (km/h)	Motor Power (Watts)	Passenger Permission	Motor Road Speed Limit (km/h)	Driving License	Mark
DK	15	yes	20	-	no	-	no	-
AT	12	-	25	600	no	-	no	-
GR	15	yes	25	-	no	50	no	-
NZ	-	-	-	300	no	-	no	-
SG	16	-	25	-	no	-	no	-
JP	16	-	20	-	no	-	no	-
ZA	-	-	25	-	no	-	no	-
TZ	-	-	-	-	no	-	no	-
RU	14	-	25	-	no	60	no	-
CN	16	yes	25	240	no	-	no	-
BJ	-	-	-	-	-	-	-	Illegal
HK	16	-	25	-	no	-	no	-
TW	16	-	25	-	no	60	yes	-
WA	16	-	16	-	no	-	no	-
NJ	17	yes	30	-	no	-	no	-
DL	16	-	25	-	no	-	yes	-
AD	16	-	20	-	no	-	no	-

with other factors. It is worth noting that although the national laws and regulations in mainland China does not set any regulations on the driving of electric scooters on the road other than minimum age-allowed and maximum speed, Beijing and Shanghai respectively banned all low-speed electric vehicles from driving on the road through local regulations in 2018 and 2016, respectively (except for electric bicycles purchased and used by individuals and legally registered and licensed). According to official explanations, these electric scooters are small in size, fast in speed, and have poor braking ability. Riders often do not wear any protective gear, which not only poses a great safety hazard to the riders, but may also cause traffic accidents and injure others, so they are completely banned from the road.

2.3.5. Stopping Sight Distance

Stopping sight distance (SSD) is known to affect safety and vehicle operations [52]. Only 5 of the 17 case studies have regulations on SSD for bicycles. Among them, HK (15–25 m) and WA (135 ft or 41.148 m) only have unified requirements for SSD, without detailed classification. AT relates SSD to bicycle speeds which are 15 m for 20 km/h, 25 m for 30 km/h, and 40 m for 40 km/h. However, as shown in Figure 8, in ZA and NJ, speed is not only linked to SSD, but also link to slope, with formulas as shown in Formula (4) [35].

$$\begin{array}{c}D={\displaystyle \frac{2.5V}{3.6}}+{\displaystyle \frac{0.5{\left(\frac{V}{3.6}\right)}^2}{2.5+\frac{9.8G}{100}}}\end{array}$$

(3)

where

• D—Stopping sight distance (meter);
• V—Bicycle design speed (km/h);
• G—Gradient (%).

However, some cyclists prefer to avoid obstacles rather than stop in front of them. Therefore, ‘decision sight distance’ is used in South Africa. Decision sight distance is defined as the distance at which a driver can, after detecting a hazard or signal, identify it or its potential threat, choose an appropriate speed and path, and perform the required actions safely and effectively [53]. The decision sight distance regulations in South Africa are shown in

Table 9. Decision about sight distances for cyclists in ZA.

Country Code	Bicycle Speed Limit (km/h)	Decision Sight Distance (m)
ZA	20	45
	30	70
	40	90
	50	115

, and the difference between stopping and decision sight distance is shown in Figure 9. The stopping sight distance is shorter than the decision sight distance, but they increase at almost the same ratio with speed.

2.4. Road and Roadside Facilities

2.4.1. Guidance Facilities

The surfacing and marking of SMI also play an integral role in the safety of facility users. The performance and smoothness of the road surface will directly affect whether users can use the facility safely. Markings separate different road users while alerting more dangerous road users (e.g., faster ones) to watch out for more vulnerable users (e.g., SMI users).

Table 10. Pavement and marking requirements of pedestrian facilities across the world.

Code	Pavement Requirement								Marking Requirement
	Flatness	Skid	Wear	Permeable	Dirt	Color	Type	Continuous	Color	Width	Sign Height	Expansion
AD	Smooth	Resistant	-	-	-	-	Firm materials	-	-	-	-	Provide
AT	-	-	-	-	-	-	-	-	-	-	≥2.0 m	-
BJ	Flat	Resistant	Resistant	√	Resistant	Mainly gray	Asphalt	-	-	-	-	-
CN	-	-	-	-	-	-	-	-	-	-	-	-
DK	-	-	-	-	-	-	-	-	-	-	-	-
DL	-	-	-	-	-	-	-	-	-	-	-	-
GR	-	-	-	-	-	-	-	-	-	-	-	-
HK	-	-	-	-	-	Special patterns	-	-	-	-	-	-
JP	-	-	-	-	-	-	-	-	-	-	-	-
NJ	-	-	-	-	-	-	-	-	-	-	-	-
NZ	-	-	-	-	-	-	Concrete and asphalt	-	-	-	-	-
RU	-	-	-	-	-	-	-	-	-	-	-	-
SG	Good	-	-	-	-	-	Concrete	-	-	-	-	-
TW	Smooth	-	-	√	-	Different from driveway	-	√	-	-	-	-
TZ	Smooth	-	-	-	-	-	-	-	-	-	-	-
WA	-	-	-	-	-	-	Concrete and bricks	-	White	4 in.	-	-
ZA	-	-	-	-	-	-	Concrete and asphalt	-	-	-	-	-

and

Table 11. Pavement and marking requirements of bicycle facilities across the world.

Code	Pavement Requirement						Marking Requirement
	Flatness	Skid	Wear	Permeable	Color	Type	Color	Width
AD	-	-	-	-	-	-	-	-
AT	-	-	-	-	-	-	-	-
BJ	Flat	Resistant	Resistant	√	Different colors only at intersections	Asphalt	-	-
CN	-	-	-	-	-	-	-	-
DK	-	-	-	-	-	-	-	-
DL	-	-	-	-	-	-	-	-
GR	-	-	-	-	Colored different	-	White	0.1 m
HK	-	-	-	-	-	-
JP	-	-	-	-	-	-
NJ	Smooth	-	-	-	-	-	White	6–8 in.
NZ	-	-	-	-	-	-	White
RU	-	-	-	-	-	-	-	-
SG	-	-	-	-	Colored	High strength materials	-	-
TW	Smooth	Resistant	-	-	Different colors may use	Different type may use	-	-
TZ	-	-	-	-	-	-	-	-
WA	-	-	-	-	Colorized not prefer	Porous asphalt	-	-
ZA	-	-	-	-	-	-	Other colors	-

are the summaries of the requirements of pavement surface and marking of pedestrian and bicycle facilities across the countries and regions analyzed. However, Table 10 and Table 11 also show us a situation that needs to be solved urgently. Many countries and regions lack road surface and road marking specifications for SMI roads. SMI users may not be able to use SMI facilities safely due to the lack of road markings or may be affected by damage caused by poor road quality, which may lead to traffic accidents.

Countries and regions that have made special regulations on bicycle lanes and sidewalks generally require bicycle lanes and sidewalks to use a different color than motor vehicle lanes. At the same time, white is the main color for markings. WA, GR, and NJ also have specific regulations on the width of markings. However, countries that have not made regulations on signs, markings, and road surfaces do not have regulations at all. Instead, they are uniformly regulated in the motor vehicle road design specifications and are not mentioned here separately.

Separating different road traffic users increases safety and facilitates traffic flows. The main methods of separation are (i) soft separation (lines and markings) and (ii) hard separation, depending on road class and specific location (green belt or barriers). As the difference between the separated objects of road traffic users increases (e.g., speed difference), the separation method between each road traffic user also changes to “harder” ones (shown in

Table 12. Separation methods across the world.

Code	Bicycle to Vehicle			Pedestrian Isolation
	Physical		Non-Physical	Physical		Non-Physical
	Barrier	Physical Divider	Buffer Strip	Column/Bollard	Curb/Low Wall	Buffer Strip
AT			√		√
NZ		√
GR			√
CN	√			√
BJ	√			√
ZA			√			√
TZ	√				√
DL				√
	Bicycles and Pedestrians
	Physical			Non-Physical
	Island	Height Separation	Curb	Line	Stripping	Colored
DK	√
GR	√	√	√
TW	√	√	√	√	√	√

).

2.4.2. Lighting

Lighting devices provide safe passage to pedestrians and cyclists as they can better see obstacles and other road users, allowing them to find their way and have a feeling of safety and security while moving [54]. Only SG and AD have clear regulations on lighting. SG requires that, on all bicycle lanes, the average lighting level should be 5 lux or higher. At the same time, the average lighting conditions at the level of an intersection should be 10 lux or higher. For a covered linkway (pedestrian path), the minimum average illuminance should be 30 lux. AD has specific regulations on the height of lighting installations. Street lighting in urban areas typically provides fixtures that should be mounted at 7.6 m to 12.2 m above grade, depending on the style of light pole and luminaire selected. But the regulations about pedestrian lighting are not such specific. AD only requires higher intensity lighting for pedestrians than for bicycles with uniform and high-density distribution of low-light lamps.

2.4.3. Weather Protections

Providing pedestrians and cyclists with shelter facilities, or shadows, against severe weather is important for safety and overall SMI attractiveness. Among the 17 cases viewed, only HK, NZ, and SG have regulations on weather protection facilities. HK encourages the installation of weather protection facilities, but they should be at least 3.3 m above ground. NZ requires weather protection to be located within the street furniture zone. However, SG has very detailed regulations on weather protection facilities which call facilities with sun and rain protection functions “covered linkway” or “covered walkway”. There are five different types or purposes of covered linkway, each with its design standards.

Table 13. Requirements of covered linkway in Singapore.

Type	Standard	HDB’s * Designs	High	Rest Area Along	Sheltered Pedestrian Waiting Area
Cross-Section Data	2.4 m × 2.4 m	3.0 m × 2.4 m	5.7 m × 5.0 m	70–100 m interval 0.8 m × 3.25 m (1.2 m × 0.9 m for wheelchair)	Provided at intersections and road crosses
Purpose and Requirements	-	Columns on one side	For cross minor roads or vehicle crossing	Keep distance to intersections

* HDB means Housing and Development Board of Singapore.

summarizes the basic design standards and requirements for covered linkway. The covered walkway uses the same requirements as the covered linkway. The covered walkway always uses the frontage of buildings as coverage and should provide lighting of at least 30 lux.

It can be concluded that developed Asian countries or regions at lower latitudes have more incentives to provide pedestrians with weather protection facilities to protect them from sun, wind, and rain. Even so, the laws, regulations, or design specifications of other countries or regions do not provide specific details about weather protection measures. Different regions may have their own local design rules, and the government may also issue government orders. This part is outside the scope of this study. However, the current extreme weather protection measures for pedestrians and cyclists provided by various countries and regions are not complete. Protecting SMI users from the sun and rain is only a basic measure. A 2015 study showed that global temperatures will continue to rise throughout the 21st century, leading to frequent natural disasters such as extreme heat, thunderstorms, and floods [55]. “Covered” lanes or paths cannot protect SMI users well enough. Other measures to keep SMI users away from extreme weather conditions should be considered.

2.4.4. Traffic Counters or Monitors

Traffic counters or monitors are extensively used to measure car volume, occupancy and speed, to manage traffic on a real-time basis or perform off-line analyses. The same applies to statistics on soft mobility traffic. However, few countries have bicycle and pedestrian counting devices. And not all countries with counting devices make the data they obtain public. Thus, in this part, we will take Paris, France as a best practice example for bicycle monitoring devices. Markedly, traffic counters are not part of design codes but are here addressed because the lack of evidence on SMI traffic is undoubtedly a major barrier to establishing evidence-based recommendations and regulations.

In 2021, Paris launched its Plan Vélo 2021–2026 (Cycling Plan 2021–2026), hoping to make Paris a “100% bikeable city”. Therefore, the city of Paris has been deploying permanent bicycle counters to assess the growth of cycling. The system Comptage vélo—Données compteurs (Bike counting—Counter data) displays and counts bicycle traffic through Eco-DISPLAY Light, a large counter developed by Eco-counter, the technology provider of the program, which can display bicycle traffic in real time (as shown in Figure 10 and Figure 11). The system can distinguish different vehicles and count different directions separately to avoid interference from other traffic participants. However, the system also has many shortcomings. For example, the system only counts bicycle traffic based on time and cannot calculate other data such as speed. At the same time, if the target lane is occupied by other traffic participants, the system will be paralyzed. Therefore, other cities should consider related issues when introducing the same project and propose solutions for these.

2.5. Supporting Facilities

Supporting facilities concern both pedestrians and bicycles and mainly include: (i) parking facilities and (ii) end-of-trip (EOT) facilities.

Bicycle parking facilities are regulated by size, facility type, and location in various regions. In TW, sizes are categorized into basic, parallel, and angled types, with specified dimensions (Figure 12). CN specifies a single bicycle size of 0.6 to 0.8 m wide and 2.0 m long, allowing diagonal arrangement in limited space without specific angle regulations. Residential areas, public buildings, and transportation hubs with high bicycle parking demand should allocate facilities according to construction guidelines and signage. Sections of roads designated for motor vehicle parking can be repurposed for bicycle parking when demand is high, ensuring accessibility near rail stations, bus stops, schools, and hospitals within 50 m. Importantly, bicycle parking facilities must not obstruct emergency passages or encroach on infrastructure like manholes or street signs.

EOT facilities cater to SMI users, such as pedestrians, cyclists, e-cyclists, or runners, providing rest, cleaning, refueling, and repair services. Although EOT facilities exist globally, they are not widespread. Singapore exemplifies this with various amenities like changing rooms, showers, and bike repair tools. Additionally, to accommodate e-bicycles and e-scooters, some companies have developed charging stations as part of EOT infrastructure.

3. Conclusions and Recommendations

As people pay more attention to the rights of pedestrians and low-speed vehicles, such as bicycles and e-scooters, the relevant urban infrastructure needs to be revised and appropriately adjusted to emerging modes. This review identified gaps and outdated design specifications, laws and regulations for SMI in many countries around the world. The absence of SMI in road design specifications has led to a lack of legal basis for SMI users. The lack of evidence in SMI design and the over-reliance and application of motor vehicle design specifications have led to poor actual use experience of SMI. SMI design did not consider actual use and was empirically developed, which led to difficulties for all road users, including traffic accidents. Therefore, in response to the current situation of SMI that is generally faced around the world, this paper selected design specifications, laws, regulations, and guidelines for SMI issued by 17 countries, regions, and major cities around the world, and conducted a horizontal comparison in terms of road surface layout, horizontal and vertical geometry, speed limits, and auxiliary facilities. At the same time, the German Guidelines for the Design of Motorways were selected for a vertical comparison between SMI and motor vehicles, which is shown in Supplementary File S2. This paper first integrated and analyzed SMI design specifications from 17 different countries or regions (Section 2). Supplementary File S2 and Section 2 use the same structure, and each section analyzes the road classification method, general layout, vertical geometry, horizontal geometry, speed, and stopping sight distance. At the same time, in each section, a summary of roadside facilities is also given. At the end of Section 2, user supporting facilities in the SMI design specification are also mentioned separately.

There are several findings about SMI in the 17 reviewed countries and regions. As for general layout (lane width, clearances, and cross-slope), except for CN, no formulaic evidence is provided to support the rationality of the data given. Basically, the data are mainly based on the grade and classification of the roads where the SMI is located. At the same time, the overall layout of some areas is basically determined by the traffic conditions of motor vehicles. As for vertical geometrics (slope and gradient), not all areas have relevant regulations, but areas with regulations are basically formulated according to the terrain of the area. As for horizontal geometrics (radii and superelevation), except for TW and HK, which provide radii, other regions do not have regulations on it. Except for TW, which provides superelevation, other regions do not have regulations on it. The data provided by TW and HK are not supported by formulas or theories. As for speeds, BJ, WA (uniform regulations), TW (based on horizontal geometry), and ZA (based on vertical geometry) regulate the speed (speed limit and/or design speed) of bicycles. The speed requirements for e-scooters in all countries and regions are not very different and are the same as the speed specified in the manufacturing standards of e-scooters, which are based on experiments. However, NZ is an exception, and they only stipulate the horsepower output. There are no regulations for the speed of pedestrians in various countries and regions. As for stopping sight distance, an equation was given by ZA and ZA also defined decision sight distance with another equation. As for the roadside facilities, guiding facilities (lane division) are mainly soft separation (marking lines) and hard separation (curb or isolation guardrail) in various countries and regions. Lighting is mainly divided into regulations on light intensity (SG for lux) and regulations on light setting height (AD). SG does the best in weather protection, with very detailed regulations. We also introduced and analyzed bicycle counting devices in France.

By comparing the SMI design specifications with those of motor vehicles, we discovered the differences between them. Firstly, the motor vehicle regulations classify lanes based on purpose, and require other road participants to adapt based on the classification of motor vehicle lanes. However, the design of SMI is based on motor vehicle roads. If the classification of motor vehicle roads cannot meet the requirements of SMI users, SMI design tends to exist independently from motor vehicle roads. Secondly, the road surface layout is mainly based on road participants with faster speeds or larger sizes, including the SMI design specifications, which are more compromised to them. Except for areas with too much SMI traffic, the situation of SMI users themselves is rarely considered when designing the road surface layout. Third, in the vertical and horizontal geometric design, motor vehicle roads will formulate different design details according to different road classifications (minimum radii of crests and sags, minimum radii of transverse curves, minimum tangent length, etc.). However, SMI simply stipulates one or several values and does not specifically refine the requirements. Fourth, the speed of motor vehicles is divided into calculating speed (i.e., design speed), recommended speed, and speed limit. Different road classifications have different speed regulations. When designing a road, the road classification and the required speed need to be determined first, and then the geometry is designed. However, in SMI design, it is just the opposite. The speed regulations of the road need to be determined according to the geometric conditions of the road. Once the conditions are not met, SMI tends to abandon the setting or set it separately from the motor vehicle lane, or even take a long detour to meet the requirements. Fifth, in terms of the requirements for stopping sight distance, the regulations of motor vehicle roads and SMI are relatively similar among these items. They both have formulas and charts to clarify. The only difference is that not all countries cite these formulas and charts in SMI specifications, resulting in the SSD regulations of some countries and regions still only giving a number to stipulate the matter, even though the formulas and charts are very clear and detailed. Finally, regarding the setting up of roadside facilities, the motorway and SMI have the same guiding facilities. Both have provisions for signs and markings. However, SMI rarely has provisions for traffic control facilities such as traffic lights, because such facilities are mainly uniformly regulated in the specifications of motorway. At the same time, due to the characteristics of SMI users, which are small in size and highly flexible, but more affected by the environment, end-of-trip service facilities (such as changing facilities, shower facilities, vehicle maintenance facilities, etc.) are relatively convenient to set up, and more countries and regions have made regulations and requirements for this.

Based on the above findings, we believe that the existing SMI design specifications, laws, and regulations should be optimized in the following aspects. First, the formula used by CN to determine the width of SMI facilities based on the traffic flow (or expected traffic flow) on the SMI facilities should be analyzed and promoted in various places. The cross-slope set based on drainage and anti-skid angles in places such as DK and ZA should also be considered. At the same time, there is currently no precedent for setting different height clearances based on the height differences of different ethnic groups in different regions, but the relationship between them and whether they should be considered is worth exploring. Similarly, the practice of setting up rest platforms or increasing lane width when the slope exceeds a certain level, as represented by WA and NJ, should be shared and promoted. However, setting different slopes according to different speeds and characteristics, and considering adding the minimum radii of the crests and sags of vertical curves, the minimum radii of horizontal curves, and the tangent length should be considered as detailed requirements in the SMI design specifications. As the speed of soft mobility vehicles increases, superelevation should not be considered a negligible factor in SMI design. Drawing on the practices of TW and motor vehicle road design, SMI specifications should make more detailed provisions for superelevation. As for speed, different vehicles can reach different maximum speeds, resulting in different safety impacts on SMI users. Referring to the specifications of motor vehicle roads, setting different speed limits for different types of roads and different stages of roads (straight lines, curves, etc.), and designing road geometry according to these speed limits should be the top priority for each country and region to optimize and improve SMI. Finally, other road design elements including SSD (such as subgrade and surface design, climbing lanes, intersection design, traffic control facilities, etc.) and traffic violation handling rules are currently almost absent or too simple. They should be supplemented or improved in full accordance with motor vehicle regulations.

Meanwhile, we need to note that environmental protection and individual experience to mobility environment are also important considerations in SMI design. The environmental and aesthetic value of SMI should be evaluated based on the actual experience of users, and user psychological feelings and landscape elements should be incorporated into SMI design to achieve the coordinated development of ecological protection, travel safety, and human experience [56,57]. Also, environmental knowledge and attitudes towards soft mobility are the most important factors affecting behavioral intentions. Local authorities should encourage people to use soft mobility more often in their daily lives through strengthening publicity and education and introducing economic incentives (e.g., using soft mobility to redeem points for discounted scenic spot tickets) [58]. E-scooters and e-bikes have obvious appeal in soft mobility systems, especially in places with good SMI such as city centers and waterfront areas. This also reflects that environmental and health benefits depend on complementarity (rather than substitution) with walking, cycling, and other mobility modes, appropriate infrastructure planning, and reuse mechanisms, otherwise it may bring negative impacts such as space occupation and resource waste [13].

Through this study, the imbalance, insufficiency, and imprecision of design regulations and standards of SMI are fully reflected. Countries that started SMI construction later are more likely to follow countries that started earlier, thereby directly borrowing their design standards. However, even in countries that started early, not all regulations are evidence-based. At the same time, almost all SMI design regulations are limited to serving bicycles and pedestrians. However, there is little coverage of medium and low-speed vehicles such as e-scooters and e-bikes that have emerged in recent years. The resilience of cycling infrastructure has emerged as a critical dimension in urban transportation systems, particularly in addressing multifaceted challenges. First, design enhancements such as physically segregated bike lanes have proven effective in reducing cyclist fatalities by 20–30% [59]. Second, lifecycle cost-effectiveness can be achieved through modular “micro-renovation” strategies, exemplified by Copenhagen’s green wave traffic lights, which boosted cycling efficiency by 40% with minimal maintenance budgets [60]. However, the study of [61] also pointed out that although e-scooters and e-bikes are included in the category of “soft mobility”, their safety is significantly different from that of traditional bicycles. The integration of soft mobility with the existing transportation system needs to be more systematic, and its position in the overall transportation system needs to be further clarified (such as replacing private cars, public transportation or walking). At the same time, there is a disconnect between soft mobility users’ safety perception of scooters and E-bikes and the actual risks, and safety education for SMI users also needs to be developed simultaneously.

This paper is the first to systematically compile and compare SMI design codes worldwide, marking a key innovation of this study. One of the aims of this study is to provide initial insights for researchers and policymakers involved in the design and implementation of SMI. By analyzing and comparing existing SMI design codes in 17 countries and regions, and comparing them with the German motor road design code, this paper brings together existing practices and expertise and identifies best practices. Crucially, the safety of soft mobility users, which is often overlooked, is highlighted. Although the analysis shows that countries around the world have made good progress in SMI design, there are still significant gaps between each other and the design of motor roads, especially in low-income and least developed countries. The findings of this paper are not country-specific but are intended to be broadly applicable. This study aims to stimulate the development of new codes and encourage countries with existing codes to improve them using evidence-based methods. Although the development of standards is often influenced by political will, the authors of this paper argue that through the contribution of this paper, even low-income and least developed countries can develop effective codes without expensive investments. As a next step, we suggest focusing on elements that currently lack empirical support to strengthen the theoretical foundation of SMI design codes. Meanwhile, targeted experiments should be conducted to address gaps in SMI design for e-scooters and e-bikes.