Development and Verification of Test Procedures for Detecting Overloading and Improper Loading in Commercial Vehicles Using a High-Speed Weigh-in-Motion System: A Case Study in Republic of Korea

Abstract

Despite continued efforts by the Korean government to improve road safety, truck-related accidents remain disproportionately fatal, with a rate approximately 2.6 times higher than that of passenger vehicles. Although legal regulations prohibit overloading and improper loading, existing enforcement practices—primarily dependent on low-speed weigh-in-motion (WIM) systems—are limited in coverage and responsiveness. This study develops and validates standardized test procedures for detecting overloading and improper loading in commercial freight vehicles using a high-speed weigh-in-motion (HS-WIM) system. The HS-WIM system offers advanced sensing capabilities, including vehicle speed, length, axle configuration, and weight measurement at highway speeds. However, Korean HS-WIM performance standards currently lack detailed guidance, especially concerning group axle load testing and asymmetric cargo detection. To address these regulatory and technical gaps, a comprehensive set of test scenarios was designed based on domestic and international standards. A dedicated testbed was constructed, and 12 commercial vehicle types were tested under varied speeds and loading conditions. The proposed procedures reliably detect violations, and the study introduces evaluation criteria that improve HS-WIM system accuracy and support future enforcement and policy development in Korea.

1. Introduction

Over the past five years, the Korean government has actively implemented a variety of traffic safety policies—such as installing additional traffic lights, deploying speed-enforcement cameras, and introducing the Safety Speed 5030 initiative. As a result, the total number of traffic accidents decreased by approximately 6.1%, from 216,335 cases in 2017 to 203,130 cases in 2021 [1].

However, truck-related accidents showed only a slight decline of 4.6% (from 27,341 to 26,081 cases) during the same period. Trucks continue to have the second-highest casualty rate after passenger cars, with a fatality rate roughly 2.6 times higher. The main causes of these accidents include speeding, overloading, and improper loading, which not only raise fatality risks but also cause significant economic losses such as cargo damage. Therefore, this study specifically targets commercial freight vehicles, where issues of overloading and improper loading are most prevalent.

To address these issues, the Korean government enforces legal restrictions on truck operations. According to the Road Act (Article 59) and its Enforcement Decree (Articles 55 and 60), trucks that exceed an axle load of 10 tons, a total weight of 40 tons, a width of 2.5 m, a height of 4 m, or 110% of the vehicle’s rated length are subject to overloading regulations [2]. Current enforcement methods, however, rely mainly on low-speed weigh-in-motion (WIM) systems that measure vehicles moving at speeds below 10 km/h, typically at checkpoints or toll booths. In mobile inspections, officers select vehicles for weighing either at random or based on visual judgment. Drivers of large trucks equipped with variable axles often manipulate axle configurations during inspections to avoid accurate weight measurement and may revert to overloaded states afterward. Such practices, combined with limited inspection coverage, result in enforcement inefficiency—the proportion of inspected vehicles on highways remains as low as 0.07–0.09% [3,4].

To improve enforcement effectiveness, the high-speed weigh-in-motion (HS-WIM) system has been proposed. The HS-WIM system provides richer data—including speed, length, vehicle classification, and weight—beyond what conventional detection or automatic vehicle classification systems can measure [5,6]. Because of its potential role in direct enforcement, system reliability is essential, and many countries have established HS-WIM performance standards to ensure measurement accuracy [7,8,9,10,11].

In Korea, the Intelligent Transportation Systems (ITS) Performance Evaluation Standards for Vehicles and Road Traffic are applied, whereas Europe uses the COST 323 standard and the United States follows ASTM E1318. Nevertheless, the Korean standard currently lacks specific provisions for group-axle load tests and offers no criteria for improper-loading detection, underscoring the need for further research in these areas [12,13,14,15,16,17,18,19].

Accordingly, this study proposes testing scenarios for evaluating the performance of HS-WIM systems in detecting both overloading and improper loading of commercial vehicles. A dedicated testbed was constructed to validate the proposed scenarios. Domestic and international regulations were analyzed to establish baseline testing conditions, and performance evaluation criteria were refined by referencing global standards. The proposed test framework—covering twelve vehicle types, driving modes, and speed parameters—serves as the foundation for defining performance evaluation criteria for HS-WIM systems, particularly in the accurate detection of overloading and improper loading.

2. Analysis of Overloading and Improper Loading Regulations and Performance Tests in Korea and Abroad

2.1. Standards for Overloading and Improper Loading Regulations in Korea and Abroad

2.1.1. Korean Regulatory Standards

In Korea, regulations governing vehicle overloading are primarily stipulated in Article 77 of the Road Act (Restrictions and Permits for Vehicle Operations), Article 79 of the Enforcement Decree of the Road Act (Vehicle Operation Restrictions), and Article 22 of the Enforcement Decree of the Road Traffic Act (Operational Safety Standards).

According to the Road Act, vehicles with an axle load exceeding 10 tons or a gross vehicle weight (including cargo) exceeding 40 tons are prohibited from operation. Additionally, the Road Traffic Act limits the permissible cargo load to within 110% of the vehicle’s rated load capacity. For example, a truck with a rated load capacity of 10 tons may legally carry a maximum of 11 tons.

These national regulatory standards are summarized in

Table 1. Overloading Control Standards in Korea.

Higher Law	Category	Detailed Information
Road Act	Article 71, Paragraph 1	Vehicles with an axle load exceeding 10 tons or a total weight exceeding 40 tons. Vehicles with a width exceeding 2.5 m, a height exceeding 4.0 m (4.2 m on designated roads), or a length exceeding 16.7 m. Vehicles that the road authority deems a particular threat to road structures or traffic safety.
Enforcement Decree of the Road Traffic Act	Article 22	The load capacity of a cargo truck must not exceed 110% of its rated load capacity. Additionally, the loaded vehicle’s dimensions are restricted to the following limits: length no more than 110% of the vehicle’s length; width within the driver’s rearward field of view (via side mirrors); and height not exceeding 4.0 m from the ground.

. Carriers intending to transport cargo beyond these prescribed limits must obtain an operation permit for restricted vehicles from the relevant authority—either the National Police Agency or the road management authority (e.g., the Ministry of Land, Infrastructure and Transport, or the respective local government).

In recent years, researchers have proposed IoT-based real-time overloading detection frameworks to supplement regulatory enforcement, enabling continuous monitoring and immediate alerting [20,21]. These approaches are expected to complement national policies by integrating real-time data collection with advanced WIM infrastructure.

2.1.2. European Regulatory Standards

The European Union follows its own regulations based on Directive 96/53/EC. As shown in

Table 2. European Cargo Truck Overloading Enforcement Standards (Tons).

Classification	Trailer (Non-Combination)	Road Train (Combination)	Articulated Vehicle (Combination)
2 axle	18	-	-
3 axle	24	-	-
4 axle		2-axle vehicle + 2-axle trailer: 36 (when axle spacing ≥ 1.3 m and <1.8 m) 2-axle vehicle + 2-axle semi-trailer: 36 (or 38) (when axle spacing ≥ 1.8 m)	2-axle vehicle + 2-axle semi-trailer: 36 (or 38) (when axle spacing ≥ 1.3 m and <1.8 m 2-axle vehicle + 2-axle semi-trailer: 36 (or 38) (when axle spacing ≥ 1.8 m)
5 axle		2-axle vehicle + 3-axle trailer: 40 (or 42)	3-axle vehicle + 2-axle semi-trailer: 40 (or 44)
6 axle		3-axle vehicle + 3-axle trailer: 40 (or 44)	3-axle vehicle + 3-axle semi-trailer: 40 (or 44)

and

Table 3. European Cargo Truck Tandem/Tridem Axle Load Standards (Tons).

Axle Spacing (d)	Tandem Axle Load	Axle Spacing (d)	Tridem Axle Load	Single Axle Load
d > 1.0 m	11	d ≥ 1.3 m	21	Non-drive axle: 10 tons Drive axle: 11.5 tons (Regardless of axle spacing)
1.0 m ≤ d < 1.3 m	16
1.3 m ≤ d < 1.8 m	18	d > 1.3 m and ≤1.4 m	24
d ≥ 1.8 m	20	d > 1.4 m

, it rigorously inspects and monitors the weight, axle load distribution, and allowable dimensions of cargo vehicles [22]. These regulations contribute to road safety, control road loads, and help reduce traffic congestion. Moreover, direct enforcement through high-speed WIM systems has been implemented in several EU member states, such as Belgium, demonstrating effective on-road monitoring and penalty automation [17].

2.1.3. U.S. Regulatory Standards

In the United States, overloading of cargo trucks is regulated by the Federal Motor Carrier Safety Regulations (FMCSRs) [23]. These regulations apply to all vehicles and drivers engaged in commercial transportation within the U.S. They include various rules to ensure road safety during transportation. As shown in

Table 4. Cargo Truck Overloading Enforcement Standards (Tons).

Classification	Federal Weight Standards	State-Specific Weight Standards
		Louisiana	California	Alabama	Minnesota
Gross Vehicle Weight (GVW)	36.29	36.29 (or 37.82, 39.92)	36.29 (or 34.84)	36.29 (or 38.1)	36.29
Single axle load	9.07	9.07 (or 8.16)	9.07 (or 8.16, 5.67)	9.07	9.07 (4.54 per single tire)
Tandem axle load	15.42	15.42 (or 15.24)	15.42 (or 15.24)	15.42 (or 16.33)	15.42

, the regulations govern the weight and size of cargo trucks, driving time limits, and other factors, all aimed at improving road safety. Recent developments in the U.S. have also introduced advanced WIM-based systems designed for automatic detection and enforcement, significantly reducing socioeconomic costs due to overweight vehicles [21].

2.2. Performance Evaluation Standards for HS-WIM Systems

2.2.1. Performance Evaluation Standards in Korea

Key performance criteria are outlined in

Table 5. Performance Standards by Evaluation Grade for HS-WIM in Korea.

Evaluation Criteria		Weight Accuracy (%)		Avoidance Driving Detection Rate (%)	Vehicle Number Recognition Rate (%)	Matching Accuracy
		Total Weight	Axle Load
Evaluation grade	Highest grade	≥95%	≥90%	≥90%	≥95%	100%
	High grade	95>, ≥93	90>, ≥85	90>, ≥80%	95>, ≥85%	100%
	Medium grade	93>, ≥90	85>, ≥80	-	85>, ≥80%	-
	Lower-medium grade	<90%	<80%	-	<80%	-

, which shows performance standards by evaluation grade for high-speed WIM systems in Korea, including weight accuracy, avoidance driving detection rate, vehicle number recognition rate, and matching accuracy.

Table 6. Number of Tests per Lane for Weight Accuracy Evaluation in Korea.

Classification	Vehicle Type	Evaluation Type
		Basic	Completion/Regular/Modification/Relocation
Number of tests	3-axle, 1-unit	10	-
	4-axle, 1-unit	10	-
	5-axle, 1-unit (with variable axle)	10	10
	5-axle, 2-unit	10	10
	Total number of tests	40	20

displays the number of test runs required per lane for weight accuracy evaluation, while

Table 7. Performance Evaluation Grades for HS-WIM in Korea.

Accuracy Grade	Utilization Plan
Class A(5)	Can be utilized for institutional purposes related to weight limits and other relevant enforcement. Can be used as a reference to verify loads measured by HS-WIM systems of grade A(5) or lower.
Class B+(7)	Can be used for enforcement in cases where the accuracy of grade A is not met, provided legal authority is granted. Can be used as a reference to verify loads measured by HS-WIM systems of grade B+(7) or lower.
Class B(10)	Collected data can be used for the design, maintenance, and evaluation of road pavements and bridge structures (e.g., to assess infrastructure aging and schedule replacements). Useful for estimating excess weight on axles and for precise vehicle classification based on load.

presents the performance evaluation grades for HS-WIM systems. Additionally, the classification of Korea’s 12 vehicle types is illustrated in

Table 8. Vehicle Types Classification in Korea.

Vehicle Class	Vehicle Type & Description	Axle Configuration (Description)
1 Class	Passenger cars and mini trucks, 2-axle unit	Small car with 2 axles
2 Class	Buses with 16+ seats, 2-axle unit	Bus with 2 axles
3 Class	Trucks under 2.5 tons, 2-axle unit	Light truck with 2 axles
4 Class	Trucks between 2.5 and 8.5 tons, 2-axle unit	Medium-sized truck with 2 axles
5 Class	3-axle unit trucks	Heavy truck with 3 axles
6 Class	4-axle unit trucks	Truck with 4 axles
7 Class	5-axle unit trucks	Large truck with 5 axles
8 Class	4-axle semi-trailers, 2-unit vehicles	Semi-trailer with 4 axles (two units)
9 Class	4-axle full trailers, 2-unit vehicles	Full trailer with 4 axles (two units)
10 Class	5-axle semi-trailers, 2-unit vehicles	Semi-trailer with 5 axles
11 Class	5-axle full trailers, 2-unit vehicles	Full trailer with 5 axles
12 Class	6-axle or more semi-trailers, 2-unit vehicles	Semi-trailer with 6 axles or more

. Emerging studies emphasize the importance of accounting for vehicle dynamic behaviors—such as wheel oscillations and axle vibrations—when calibrating high-speed WIM sensors, as these significantly influence accuracy under real-world conditions [8,24].

2.2.2. Performance Evaluation Standards in Europe

The performance evaluation standards for high-speed WIM systems in Europe follow the COST 323 standard. This standard was developed under the COST program to measure vehicle weight in real time on roads, enabling overloading detection and road load monitoring to improve safety. Key details are shown in

Table 9. Performance Evaluation Standards for HS-WIM in Europe.

Classification	Accuracy Grade: Allowable Range of Confidence Interval δ (%)
	A(5)	B+(7)	B(10)	C(15)	D+(20)	D(25)	E
Total weight	5	7	10	15	20	25	>25
Combined axle	7	10	13	18	23	28	>28
Single axle	8	11	15	20	25	30	>30
Single axle within a combined axle	10	14	20	25	30	35	>35
Speed	2	3	4	6	8	10	>10
Axle distance (Axle spacing)	2	3	4	6	8	10	>10
Total flow	1	1	1	3	4	5	>5

. One main point is that dynamic impacts between the road surface and vehicles affect measurement accuracy. Therefore, the standard considers not only total weight and axle load but also combined axle loads and single axle loads within a combination. The standard categorizes accuracy into seven levels based on allowable tolerances and recommends accuracy levels of A(5) or B+(7) for HS-WIM systems suitable for overloading enforcement.

2.2.3. Performance Evaluation Standards in the United States

The U.S. performance evaluation standards for HS-WIM systems primarily follow ASTM E1318. This standard includes requirements for accuracy classes (such as Type I for enforcement) and prescribes testing methods under various conditions. For example, the California Department of Transportation requires WIM systems to meet pavement smoothness criteria (±3 mm over 12 m) as specified in ASTM E1318. Based on such practices, to ensure reliable enforcement use, the HS-WIM systems must demonstrate high accuracy across different vehicle types and operating conditions.

Type I: A lower-accuracy category (±25% wheel load, ±20% axle load, ±10% gross weight) meant for screening.

Type II: Intermediate, not commonly used in enforcement.

Type III: The standard for enforcement-level systems (e.g., ±6% gross weight, ±10% combined axle weight tolerance).

Type IV: High-precision slow-speed or static systems (nearly equivalent to static scale accuracy).

3. Methodology

This section outlines the methodology used to derive test scenarios, design the HS-WIM testbed, configure sensors, and evaluate data, thereby establishing a structured procedure for detecting overloading and improper loading in commercial vehicles.

3.1. Derivation of Overloading Test Scenarios

A key insight from the international comparison is that Korea’s current ITS performance standards lack provisions for group-axle load assessment and improper loading evaluation. For example, as shown in

Table 10. Current Evaluation Criteria for High-Speed Weigh-in-Motion (HS-WIM) Systems in Korea and Abroad. (○ = applicable, × = not included in criteria.)

Evaluation Item	Korea ITS	COST 323	OIML R134	ASTM E1318
Total weight measurement	○	○	○	○
Wheel (single axle) weight	×	×	×	○
Axle load measurement	○	○	○	○
Group-axle load measurement	×	○	○	○
Single axle within group	×	○	×	×

, Korean standards (prior to this study) did not include a category for group-axle load tests (indicated by “×” in the table under combined axle load), whereas international standards like COST 323, OIML R134, and ASTM E1318 do address multi-axle group weights. Additionally, while Korea recently defined basic criteria for improper loading (e.g., legal height, length, width limits), there were no standardized test methods to evaluate a WIM system’s ability to detect such violations. This necessitated developing new evaluation methods for improper loading detection.

Domestic and international WIM standards were reviewed to derive realistic overloading test scenarios for HS-WIM systems. For enforcement purposes, scenarios were defined to challenge the system’s accuracy under varying loads and speeds. Twelve overloading scenarios were established, reflecting combinations of load levels (from no load up to 110% of rated load), vehicle speeds (e.g., 50, 70, 90 km/h), and driving patterns (centered vs. deliberate lateral offset driving). These scenarios also include different environmental conditions (day vs. night, dry vs. wet pavement) to evaluate system robustness.

The overall process for implementing these test scenarios is illustrated in Figure 1, which outlines the overload vehicle detection test procedure.

The overloading test conditions were defined in compliance with legal thresholds. For instance, scenarios include vehicles loaded to exactly 100% and 110% of their legal load limit to test detection of legal vs. illegal loading. Each scenario was repeated multiple times (typically 10 runs each) to ensure statistical reliability.

Table 11. Overloading Condition Test Scenarios.

Parameter	Condition
Load weight	0%, 50%, 100%, 110%
Speed	50 km/h, 70 km/h, 90 km/h
Driving pattern	Normal driving (centered), Eccentric driving (driver-side wheel on the edge)
Weather	Normal, Snowfall, Rainfall
Group driving	Single vehicle, Same vehicle type, Combination vehicle type
Number of repetitions	10 times

summarizes the overloading condition test scenarios.

3.2. Derivation of Improper Loading Test Scenarios

There are no existing standardized test conditions for improper loading of cargo vehicles in Korea or abroad. Therefore, we developed test scenarios based on Korea’s regulations on improper loading, as shown in

Table 12. Improper Loading Condition Test Scenarios.

Parameter	Condition
Load type	Height exceeding 4.0 m, Length exceeding 16.7 m, Width exceeding 2.5 m
Speed	50 km/h, 70 km/h, 90 km/h
Driving pattern	Normal driving (centered), Eccentric driving (driver-side wheel on the edge), Deviation driving (driver-side wheel off the lane)
Sunlight conditions	Daytime
Group driving	Single vehicle, Same vehicle type, Combination vehicle type
Number of repetitions	10 times

. The scenarios address factors such as load shape (oversize dimensions), vehicle speed, driving patterns, sunlight conditions, group driving, and number of repetitions. To verify these scenarios, real vehicle tests were conducted on 12 vehicle types under intentionally improper loading conditions.

Based on these scenarios, an improper loading model was created (see Figure 2) to conduct the tests. The tests followed procedures outlined in Figure 3. Evaluation criteria included vehicle number recognition rate and specification measurement accuracy.

For the improper loading model, we defined quantitative thresholds: a cargo height exceeding 4.0 m, a cargo length exceeding 16.7 m, and a cargo width exceeding 2.5 m. To simulate real-world unbalanced loading, the cargo’s center of gravity in the model is intentionally offset from the vehicle’s centerline. This ensures that the test scenarios adequately represent vehicles with improperly distributed loads.

3.3. Testbed Design and Sensor Configuration

To verify the test scenarios, a testbed was constructed on the Saemangeum Seawall in Gunsan, where vehicle traffic could be controlled and high-speed driving tests could be conducted, as shown in Figure 4 The overloading detection system installed on the testbed consists of three main components: the WIM system, the AVI (Automatic Vehicle Identification) system, and the VPS (Visual Positioning System). The WIM system is embedded in the road surface and is responsible for measuring the total weight and axle load of passing cargo vehicles. The AVI system uses high-resolution cameras to capture images of all passing vehicles and recognizes their license plates. The VPS measures the dimensions (height, width, length) of the cargo trucks, including the shape and size of the load. The information measured by the installed systems is summarized in

Table 13. Specification Data Measurable by Vehicle Specification Measurement Sensor Types.

Sensor Type	Classification	Detailed Specification Name
Road-embedded Contact Sensors	Weight Information	Axle load, total weight, combined axle weight, wheel load
	Axle Information	Number of axles, axle configuration, wheel type
	Vehicle Information	Vehicle type
	Length Information	Vehicle length, axle spacing, wheelbase, overhang
	Driving Information	Speed, acceleration, driving pattern (partial)
Non-contact Sensors Installed on Pole Structures	Vehicle Information	Vehicle number
	Driving Information	Driving pattern
	Volume Information	Vehicle length, width, height

.

4. Test Results

4.1. Overloading Test Results

Using the HS-WIM system installed at the testbed (Figure 5, overloading tests were conducted on cargo vehicles. Real vehicle driving tests were performed to assess basic performance, and indicators such as weight accuracy (total weight and axle load), avoidance driving detection rate, matching accuracy, vehicle number recognition rate, and specification measurement accuracy were evaluated. The results were analyzed by applying the performance evaluation standards outlined in the ITS Performance Evaluation Standards for the Automotive and Road Traffic Sectors.

Additionally, overloading tests considering group driving conditions (Figure 6 and overloading tests under rain and snow conditions (Figure 7 were conducted to evaluate HS-WIM performance under various conditions.

Representative results are shown in

Table 14. Summary of HS-WIM Performance Test Results for 25-ton Cargo Truck.

Evaluation Criteria	Classification Total Weight (Gross)	Details Driving (Times)	1-Unit Vehicle			Accuracy (%)
			50 km	70 km	90 km
Weight accuracy	1st axle load (AOW) 2nd axle load (AOW) 3rd axle load (AOW)	Max (%)	10	10	10	98.7
		Min (%)	0.5	1.1	1.1
		Driving (times)	−0.8	0.2	−0.4
	4th axle load (AOW) Vehicle number recognition Matching accuracy	Max (%)	10	10	10	96.6
		Min (%)	−0.9	1.4	4.1
		Driving (times)	−3.4	−1.1	−1.0
	Avoidance driving detection Classification	Max (%)	10	10	10	97.6
		Min (%)	1.1	2.3	1.4
		Driving (times)	−0.7	−0.4	−1.3
	Total weight (Gross) 1st axle load (AOW) 2nd axle load (AOW)	Max (%)	10	10	10	97.3
		Min (%)	1.1	−0.1	2.8
		Driving (times)	−3.2	1.1	−2.1
	3rd axle load (AOW) 4th axle load (AOW) Vehicle number recognition	Max (%)	10	10	10	96.9
		Min (%)	1.9	0.5	2.2
		Driving (times)	−3.0	−0.5	−1.2
Vehicle number recognition Avoidance driving detection (Deviation) Vehicle number recognition	Matching accuracy Avoidance driving detection	Detection (times)	10	10	10	100
		Driving (times)	10	10	10
	Specification measurement accuracy Classification	Mismatch (times)	10	10	10	100
		Driving (times)	0	0	0
Avoidance driving detection (Deviation)	Total weight (Gross) 1st axle load (AOW)	Detection (times)	10	10	10	100
		Driving (times)	10	10	10
Vehicle number recognition	2nd axle load (AOW)	Detection (times)	10	10	10	100
		Details	10	10	10

, which summarizes performance evaluation outcomes for a 25-ton cargo truck (7-axle, 5-unit vehicle) under conditions such as 110% load, speeds of 50/70/90 km/h, normal vs. nighttime driving, fair weather, and single-vehicle operation. For weight accuracy, all axles showed over 95% accuracy, achieving the highest grade. For avoidance driving detection, all 10 deliberate lane deviations were correctly identified (100% detection rate). The vehicle number recognition rate was 100%, with all 10 passes correctly recognized. Matching accuracy also reached 100%, as the test vehicle’s number matched the recognized number in all cases. The specification measurement accuracy was 100% for all measured dimensions.

For a more detailed analysis of axle load errors, the measurements are depicted in Figure 8, Figure 9, Figure 10, Figure 11 and Figure 12. Figure 8 presents the axle load error rate for the first axle under different speed conditions. The maximum error was 3.4% at 50 km/h and 4.3% at 90 km/h, an increase of 0.9% attributed to greater vehicle body vibration at higher speed. Figure 9 shows the second axle load error rate, which remained within 2.5% across all speeds. Figure 10 shows the third axle load error rate: 3.2% at 50 km/h, 1.2% at 70 km/h, and 2.8% at 90 km/h. The dip in error at 70 km/h is believed to result from suspension characteristics at that speed. Figure 11 compares the standard vs. measured axle loads for each axle, and Figure 12 presents the overall axle load error rates. Overall accuracy was 98.4% for the first axle, 99.2% for the second axle, 98.7% for the third axle, and 99.0% for the fourth axle. The total vehicle weight accuracy was 98.8%, corresponding to the highest performance level.

Based on these results, the effectiveness of the overloading test scenarios using HS-WIM systems for commercial vehicles was confirmed. The scenarios are expected to be highly applicable across various HS-WIM implementations.

4.2. Improper Loading Test Results

Using the HS-WIM system (Figure 5), improper loading tests were conducted on cargo vehicles. As shown in Figure 13, the system measured the specifications of cargo trucks, including the shape and size of the load (height, width, length). To evaluate improper loading detection under various conditions, tests were performed during both day and night, and with single vs. group vehicle scenarios, as shown in Figure 14.

Representative results from the improper loading tests are summarized in

Table 15. Summary of Improper Loading Detection Performance Evaluation Test Results (7-class, 5-axle, 1-unit Vehicle).

Evaluation Criteria Improper Loading Detection	Classification Height Exceeding 4.0 m	Details Driving (Times)	5-Class, 3-Axle, 1-Unit Vehicle			Accuracy (%)
			50 km	70 km	90 km
Vehicle number recognition Evaluation criteria Improper loading Detection Vehicle number recognition	Length exceeding 16.7 m Width exceeding 2.5 m	Detection rate(%)	10	10	10	100
		Driving (times)	100	100	100
	Vehicle number recognition Matching accuracy	Detection rate (%)	10	10	10	100
		Driving (times)	100	100	100
	Avoidance driving detection Classification	Detection rate (%)	10	10	10	100
		Driving (times)	100	100	100
Evaluation criteria Improper loading Detection Vehicle number recognition	Height exceeding 4.0 m Length exceeding 16.7 m	Detection (times)	10	10	10	100
		Driving (times)	10	10	10
	Width exceeding 2.5 m Vehicle number recognition	Mismatch (times)	10	10	10	100
		Driving (times)	0	0	0
Evaluation criteria	Matching accuracy	Detection (times)	10	10	10	100
		Details	10	10	10

. The table provides the performance outcomes for detecting improperly loaded vehicles (e.g., over-height or off-center loads). In these tests, the HS-WIM system correctly identified 100% of instances where loads exceeded legal height or width limits, and it successfully flagged all cases of significant load offset on the vehicle bed. Vehicle identification (plate recognition) remained at 100%, and there were no false positives (vehicles incorrectly flagged as improperly loaded when they were within legal limits).

5. Conclusions

This study developed and validated standardized test procedures for detecting overloading and improper loading in commercial vehicles using a high-speed WIM system, using Korea as a case study. Through designing a dedicated testbed and conducting comprehensive experiments, the following contributions were made:

Establishment of standardized HS-WIM test procedures: We proposed a set of test scenarios and procedures to evaluate HS-WIM system performance in detecting both overloaded and improperly loaded vehicles. These scenarios cover a wide range of conditions (load levels, speeds, driving patterns, weather) and can serve as a framework for evaluating WIM systems in enforcement applications.

Construction of a validated HS-WIM testbed under real-world conditions: We built and utilized a full-scale testbed incorporating road-embedded WIM sensors, AVI cameras, and VPS sensors. The testbed allowed us to simulate enforcement conditions and validate the HS-WIM system’s accuracy and reliability in a controlled yet realistic environment.

Performance evaluation criteria for regulatory development: We proposed quantitative performance criteria (e.g., required accuracy levels, detection rates) based on both domestic standards and international best practices. The findings and criteria from this study can inform national regulations and guidelines for deploying HS-WIM systems in direct enforcement.

Overall, the results demonstrate that a high-speed WIM system, when evaluated and calibrated through the procedures we developed, can reliably identify overloaded and improperly loaded commercial vehicles. The new test procedures and performance metrics introduced in this work provide a foundation for improving enforcement of truck loading regulations, not only in Korea but also in other jurisdictions aiming to adopt advanced WIM technologies for road safety and infrastructure protection.

Future Work: While the developed test scenarios and criteria effectively addressed current gaps (such as group-axle loads and improper loading detection), future studies could explore additional scenarios (e.g., extreme weather or road conditions) and further optimization of sensor fusion algorithms. Collaborative efforts between industry and regulatory bodies will also be important to translate these findings into updated standards and widespread implementation of HS-WIM systems for enhanced enforcement and safety.