Towards Smart and Sustainable Last Mile Delivery Systems: A Scoping Review and Conceptual Framework

Abstract

The accelerated growth of e-commerce and ongoing urban expansion have intensified the challenges associated with last-mile delivery, making it a critical issue in sustainable urban logistics. Therefore, our paper presents a scoping review to systematically delineate the current state of research on smart and sustainable last-mile delivery systems. We explore both innovative technologies—such as artificial intelligence, autonomous vehicles, the Internet of Things, and digital twins—and human-centered dimensions, including urban design, policy development, and collaborative stakeholder engagement. Using the PRISMA-ScR-based methodology, 140 peer-reviewed articles (2015–2025) have been analyzed to highlight key trends, gaps, and prospective directions. The study underlines how the technologies of Industry 4.0 have improved visibility and operational efficiency, but holistic thinking that incorporates environmental, human, and policy factors remains undeveloped. Based on these findings, this article provides a conceptual framework for smart and sustainable last-mile delivery, focusing on the intersection of digital and simulation tools and human-centric governance to achieve optimized efficiency, environmental performance, and equity. This framework helps both academics and decision-makers to advance data-driven, resilient, and integrative city logistic ecosystems.

1. Introduction

Urban logistics has become a major challenge for modern cities, as the continued growth of e-commerce and on-demand services intensifies pressure on transportation networks and urban sustainability [1]. Last-mile delivery refers to the final stage of moving goods from a distribution node to the end customer and has become the costliest and most emission-intensive segment of urban supply chains [2]. According to a study published by MDPI, the average cost per delivery is USD 10.1, while customers are charged an average of USD 8.1, which significantly affects profit margins [3]. Urban density therefore tends to increase last-mile delivery costs, which often exceed 50% due to congestion, the high number of stops, and access difficulties [4]; also increased are local emissions, which account for nearly 30% of freight-related CO₂ emissions in densely populated urban areas [5]. As a result, improving the sustainability of LMD systems has become a strategic priority for cities pursuing climate neutrality and smart mobility goals under the Sustainable Development Goals (SDGs) [6,7,8].

The latest Industry 4.0 advances and evolving Industry 5.0 paradigms have created significant opportunities for revolutionizing last-mile delivery [9]. Emerging technologies such as AI, IoT, autonomous vehicles, and digital twins offer real-time optimization, the possibility of predicting outcomes, and data-based decision-making [10,11,12]. Tools like the Simulation of Urban MObility (SUMO) allow researchers and schedulers to test how different delivery strategies affect the environment, space, and time [13,14,15]. Combined, these tools pave the way for a smarter delivery ecosystem able to reduce emissions, boost service reliability, and optimize city infrastructure use.

However, even with all these advanced technologies, building a sustainable last-mile delivery system needs more than just digital optimization [2,9]. Success also depends on people and governance—how stakeholders cooperate, how public policy aligns, and whether communities accept and benefit from these systems [16]. Many existing studies focus on the technological or cost-effectiveness aspect [3,10,17,18], while few studies have considered the environmental and social dimensions [6,12,13,19]. This unbalanced landscape has limited the inclusivity and scalability of intelligent logistics solutions.

To move forward, cities need integrated frameworks that link smart technologies and sustainable principles—combining data-based intelligence with collaborative urban logistics scheduling for developing flexible, equitable, and adaptable delivery systems [18,19]. In addition, simulation-driven approaches, such as SUMO (1.25.0 version), can provide data-based testing and refine environments for these integrated solutions before their real deployment [15].

Therefore, this article seeks to map and systematically synthesize current research related to smart and sustainable last-mile delivery systems using a scoping review methodology, and to provide a conceptual framework that interlinks technological advancement, the imperatives of sustainability, and human-centric governance. This study aims to highlight key technological and sustainable trends driving research on LMD and how they are employed for modeling LMD systems to develop a comprehensive framework combining environmental, digital, and social dimensions for intelligent and sustainable urban delivery solutions.

The research paper is organized into five sections. After the introduction, Section 2 outlines the methodology and describes the PRISMA-ScR-based process. Section 3 synthesizes the findings, the thematic mapping, and the main challenges related to building SSLMD systems. In Section 4, we propose the conceptual framework and outline its implications. Finally, in Section 5, we draw conclusions and provide guidance for future research and policy directions.

2. Methodology

2.1. Research Design

2.1.1. The Significance of the Method

To conduct a literature review effectively and appropriately, several methods can be used, such as the PRISMA, bibliometric analysis, scoping review, and meta-analysis methods.

Table 1. Comparison of literature review methods applied to last-mile delivery research.

Method	Objective	Data Type	Transparency	Use Case	Ref
Systematic Review	Answer a specific research question with critical synthesis	Qualitative and/or quantitative	Very high	When evaluating the effectiveness of solutions or strategies	[6,7,8,20,21]
Scoping Review	Investigate wide-ranging and quickly evolving fields	Qualitative and/or quantitative	Very high	When concepts are still developing	[19,22,23]
Meta-Analysis	Quantitatively combine results from multiple studies	Quantitative only	Very high	When comparing intervention outcomes	[23,24]
Bibliometric Analysis	Analyze trends, networks, and influential publications	Bibliographic metadata only	Moderate	When identifying key authors, keywords, and topic trends	[25,26]

provides a comparison of these methods based on various evaluation criteria. Given the complexity and interdisciplinary nature of our research, a scoping review methodology was adopted in accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Scoping Reviews, Table S1) process. Distinct from systematic reviews, scoping reviews are especially relevant for investigating wide-ranging and quickly evolving fields in which concepts are still developing, such as smart and sustainable last-mile delivery.

2.1.2. Research Questions

The study has addressed the following key research questions:

(1)

What are the most important sustainable and technological trends modeling last-mile delivery systems?

(2)

In what ways are digital and simulation-driven technologies (e.g., SUMO, AI, IoT, and digital twins) being incorporated within LMD research?

(3)

How can conceptual and systematic framework gaps be addressed to contribute towards developing a unified framework for SSLMD?

2.2. Data Collection and Screening

The literature search was conducted using three academic databases: Scopus, Web of Science (Wos), and Dimensions, carefully selected for their exhaustive coverage of the field of transport, sustainable development, and technology-focused research. Therefore, a set of relevant keywords was identified and developed iteratively, combining Boolean operators such as (“urban logistics” OR “last mile delivery” OR “city logistics” OR “last mile transportation”) AND (“green logistics” OR “sustainability” OR “social impact” OR “decarbonization” OR “cost efficiency” “social impact”) AND (“intelligent” OR “smart” OR “digital” OR “IoT” OR “AI”OR “autonomous vehicles” OR “simulation”).

The searching path was applied across all three databases for the period of 2015–2025. For each database, we have developed a reproductible set of Boolean search strings. According to the PRISMA-ScR guidelines, we have provided the exact queries, the fields, the filters, and the number of records retrieved in

Table 2. All-inclusive search strategy used in the databases.

Database	Search String (Exact Query Used)	Fields	Filters	Results
Scopus	TITLE-ABS-KEY(“last mile deliver *” OR “last-mile deliver *” OR “urban logistics” OR “urban freight” OR “city logistics”) AND TITLE-ABS-KEY(“smart cit *” OR “smart logistics” OR “intelligent transportation” OR “IoT” OR “Internet of Things” OR “AI” OR “artificial intelligence” OR “digital twin *” OR “autonomous deliver *” OR “robotic deliver *”) AND TITLE-ABS-KEY(“sustainab *” OR “green logistics” OR “environmental impact” OR “low emission *” OR “decarbon *” OR “energy efficien *”)	Title, Abstract, Keywords	English; Article + Conference Paper	302
Web of science	TS = ((“last mile deliver *” OR “last-mile deliver *” OR “urban logistics” OR “city logistics” OR “urban freight”) AND (“smart cit *” OR “smart logistics” OR “intelligent transportation system *” OR “ITS” OR “IoT” OR “Internet of Things” OR “AI” OR “artificial intelligence” OR “digital twin *” OR “autonomous vehic *” OR “autonomous deliver *”) AND (“sustainab *” OR “green logistics” OR “low carbon” OR “decarbon *” OR “energy efficien *” OR “environmental impact”))	Topic (Title, Abstract, Author Keywords, Keywords Plus)	English;Article + Proceedings Paper	181
Dimensions	(“last-mile delivery” OR “last mile logistics” OR “urban freight” OR “urban logistics” OR “city logistics” OR “sustainable logistics” OR “green logistics” OR “smart logistics”) AND (“AI” OR “artificial intelligence” OR “machine learning” OR “IoT” OR “industry 4.0” OR “industry 5.0” OR “autonomous vehicles” OR “digital twins” OR “smart city” OR “intelligent transport systems”)	Title, Abstract, Concepts, FOR codes	English; Articles, Conference proceedings.	43

.

After reviewing the data, 92% of the articles from Dimensions were already indexed in Scopus, and the remaining 8% were indexed in Wos. So as to avoid any redundancy in our literature review, we decided to use only Scopus and Web of Science for the rest of our analysis. We used a careful filtering process which involved the following inclusion criteria:

• Peer-reviewed papers published in English between 2015 and 2025;
• Papers explicitly studying LMD in urban logistics;
• Papers providing empirical, conceptual, or simulation frameworks.

The exclusion criteria included were as follows:

• Studies on non-urban logistics or non-urban freight transport;
• Non-English publications;
• Studies not relevant to sustainability or smart technologies.

As illustrated in Figure 1, 526 publications were identified at first. After screening, 198 papers were retained. Regarding eligibility, the full-text assessment identified 140 papers that met the inclusion criteria.

3. Results and Discussions

3.1. Data Extraction and Categorization

After the inclusion and eligibility phases, all selected papers (n = 140) were imported into Nivo for coding and synthesis. The data extraction and coding process was based on the standardized steps recommended by Tricco et al. [23] for scoping reviews.

All publications included were reviewed in their entirety, and the relevant variables were extracted: bibliographic data; objective and research questions; method used; technology used; sustainability dimensions; and key findings.

As a result, Figure 2 shows the growing interest in the topic of sustainable last-mile delivery, which has continued to increase over the years. Between 2015 and 2019, there were fewer than six publications per year. However, from 2020 onwards, we see a significant increase: publications indexed in Wos reached 16 in 2021 and rose to 28 in 2025, while those indexed in Scopus rose from 12 to 30 over the same period. This trend reflects the growing recognition of last-mile logistics as a crucial issue in urban logistics, particularly in the face of environmental challenges and the rise of e-commerce.

In addition, collectively, the studies represent an expanding base of knowledge on SSLMD. This research base covers several regions, concentrated in Europe (42%), Asia (27%), and North America (20%), with the other 11% coming from emerging economies.

Additionally, the pattern of publication shows a dominance of articles in peer-reviewed journals (78%), then conference papers (16%), and book chapters or reports (6%). The trend of consistently increasing publications over the last decade underlines a growing interest in the digitization and sustainability of last-mile logistics, driven by the emergence of Industry 4.0/5.0 technologies and sustainable urban policies.

Regarding the VOSviewer-generated bibliometric map in Figure 3, which clearly and accurately shows how research on sustainable last-mile delivery has evolved over the past decade, we can see at the core of the map major themes such as last-mile delivery, sustainability, and vehicle routing problems, highlighting how the optimization of urban logistics has become central to research discussions. Around these key concepts, we find significant environmental concerns, including carbon emissions, energy efficiency, and green logistics, which clearly show that sustainability is now a main topic of discussion. One issue really stands out: urban congestion. We are also seeing exciting new topics emerge, such as autonomous vehicles, machine learning, and bottom-up electrification, all of which signal a shift toward smarter, cleaner, and more responsive delivery systems.

To perform categorical coding and cross-dimensional analysis, the data was first compiled in Microsoft Excel, before being imported into NVivo. All publications were classified based on defined categories from the SSLMD framework, as follows: technological, sustainability, and human-centric dimensions. The categorization summary of the included studies by research dimension is presented in

Table 3. Synthesis of the categories of the included studies by the three main dimensions of SSLMD.

Main Dimension	Categories	Focus/ Technology	Nb Studies	Ref
Technological	Artificial Intelligence and Machine Learning	Dynamic routing, anticipatory demand, optimization methodologies	31	[9,10,11,12]
	IoT and Sensor Networks	Real-time tracking, intelligent storage solutions, connected vehicles	22	[19,27,28]
	Digital Twins and Simulation	AnyLogic, multi-agent systems, SUMO, digital twins for logistics	17	[15,23]
Sustainability	Environmental Efficiency	Reduction in CO2 emission, eco-routing, green vehicles	21	[2,3,4,5,6,7,8,13,17,18,20,24]
	Cost Efficiency	Sharing logistics, cost optimization, resource allocation	15
	Social Sustainability	User accessibility, acceptance, equity, and inclusivity	14
Human-Centric	Regulation and Policy	Data governance, public–private partnerships, urban freight policies	9	[16]
	Stakeholder Collaboration	Collaborative design, logistics integration, sharing infrastructure	7	[21,22]
	Risk Management and Resilience	Disruption management, adapting to climate change, redundancy strategies	4	[29,30]
Total			140

.

After this, the codification process was carried out using a combination of inductive and deductive methods: a deductive method was based on the three dimensions already illustrated, and the inductive method was based on emergent issues, such as shared logistics, micro-hubs in cities, e-governance, acceptance, and resilience, identified from the scoping review.

In order to increase methodological consistency, we have produced a codebook. Each study was coded under one or more themes. For each one, we have included a brief definition, inclusion and exclusion criteria, rules for deciding ambiguous cases, illustrative textual examples, and cross-references to related codes. The codebook was revised across three iterative cycles and was finalized prior to the commencement of full coding. The complete codebook is provided in Appendix A for transparency and reproducibility purposes. Therefore, two researchers cross-validated 50 (≈ 36%) randomly selected files. Also, inter-rater reliability was calculated using Cohen’s Kappa coefficient (K = 0.82), indicating substantial agreement.

3.2. Data Analysis and Synthesis

The results of the codification process visualize, using the NVivo network and co-occurrence matrices, the relationship between technological and sustainability issues. Figure 4 highlights the thematic categorizations and analytical dimensions. These outcomes were used to develop an empirical basis for the integration scenarios that were subsequently formalized within the conceptual framework.

The generation of co-occurrence matrices was carried out in order to quantify the frequency with which two themes appeared together in all of the studies. Each cell in the matrix represents the corresponding number of coded sources across both nodes, allowing us to identify “flashpoints” within the research landscape. The results showed the following (Figure 5):

• A strong co-occurrence between innovation technologies and sustainability, signaling that most of the research on LMDs focus on efficiency and emission reduction.
• A modest co-occurrence between human-centered issues and digital technologies, indicating the limited integration of social dimensions.
• A lack of co-occurrence between policies/regulations and digital twin/simulation-driven approaches, suggesting that the data-based testing of policies is understudied in the current literature.

In brief, the heatmap highlights the necessity for multidimensional frameworks that integrate human and political insights into technology-enhanced delivery systems. These empirically based co-occurrence models have served in the building of the conceptual framework for SSLMD proposed in Section 4.

3.3. Discussions: Challenges and Opportunities

The scoping review and synthesis indicate that the roadmap to smart and sustainable last-mile delivery systems is both ambitious and complex. This sub-section discusses the key challenges and opportunities organized into four thematic areas: technological integration, sustainability, governance and policy complexity, and human-centric transformation.

• Technological integration: One of the main challenges identified in this study is the segmented nature of digital technology deployment in city logistics networks. Despite extensive research on AI, IoT, and simulation tools, these technologies still work in isolated ways and not as part of an interconnected data ecosystem. The current research focuses on the optimization of special items (e.g., distribution modes and fleet planning) without taking into account the systemic connectivity between data flows, urban infrastructure, and logistics actors [10,11,18]. There are opportunities in exploiting digital twin environments to integrate different data sources [31]. Also, AI-driven analytics and machine learning optimization offer the important potential for collaborative decision-making, particularly once coupled with cutting-edge computing for decentralized intelligent delivery operations [32,33].
• Sustainability: The second key challenge is finding trade-offs in sustainable LMD. Although many studies focus on environmental performance, like reducing CO₂ emissions and electrifying vehicles, few studies consider economic feasibility or social inclusivity [34]. Thus, the building of integrated sustainability assessment frameworks that balance the three dimensions offers new opportunities. Additionally, approaches such as multicriteria decision-making and life-cycle sustainability assessments can offer integrated evaluations of innovations like drone deliveries and crowdshipping models [35].
• Complexity of governance and policy: This study outlines an important gap between the speed of technological advances and the slowness of policy adaptation [36]. Current urban freight regulations often remain behind developments such as self-driving delivery vehicles, algorithm-based vehicle routing, and platforms for data sharing. This gap generates uncertain conditions for regulation and restricts the evolution of sustainable solutions [36]. There are opportunities in adaptive governance models, in which policies develop dynamically through data feedback loops generated by intelligent systems [37].
• Human-centric transformation: Despite the growing use of automation, the human-centric aspect is still the least explored part of research on SSLMD. There are few studies examining the way end users, riders, and citizens experience and deal with the changing technology in delivery systems [38]. However, the current transition to Industry 5.0 offers an opportunity to refocus on people in the innovative process [39].

Therefore, four main gaps in the research were identified through the scoping review:

• Fragmentation in the integration of the smart and sustainable aspects: most studies focus on technology or sustainability separately.
• Limitations in inter-field modeling: only a few models combine simulation results with socioeconomic parameters.
• Lack of policy-oriented frameworks: often, urban logistics evolve more rapidly than policy.
• Unexplored human-centered evaluation measures, such as the perceivability of fairness, inclusiveness, and accessibility.

These gaps motivate the elaboration of a conceptual framework linking the technological, environmental, and human dimensions within LMD systems.

While this study consolidates data published between 2015 and 15 October 2025, we are aware that innovations in AI, robotics, and smart logistics continue to evolve rapidly. As such, the findings should not be viewed as a definitive account of current trends, but as a reflection of the knowledge available at the time of writing. Ongoing research, which will include developments in late 2025 and early 2026, are essential to remain aligned with emerging technology advancements.

4. Conceptual Framework for SSLMD

4.1. Framework Overview

Based on the results of the NVivo thematic synthesis, co-occurrence analysis, and PRISMA-ScR of 140 research studies, a conceptual framework was designed to highlight the relationships underlying the development of smart and sustainable last-mile delivery systems. This framework (see Figure 6) incorporates three fundamental dimensions: human-centered governance and the integration of sustainability and technological innovation, linked together by retroactive loops that improve the intelligence, resilience, and inclusivity of the system. The proposed framework employs three interrelated layers:

• Technological innovation layer: This represents the digital infrastructure that enables the system’s intelligence. It incorporates technologies such as artificial intelligence, blockchain, the Internet of Things, and digital twins, allowing for advanced optimization, data analytics, and self-governed decision-making. Such technologies jointly improve the efficiency of routing and cargo sharing and reduce emissions, thereby building the technical groundwork for intelligent logistics operations.
• Sustainability integration layer: The intermediate layer focuses on the three dimensions: environmental protection, economic efficiency, and equity. It is in alignment with the United Nations’ Sustainable Development Goals (SDGs 9, 11, and 12) and the European Green Deal [40] by encouraging low-carbon, cost-effective, and responsible delivery models. The main means involve assessing eco-efficiency, adopting renewable energy, electrifying vehicles, and implementing reverse logistics, which guarantee that the operational enhancements will contribute to a wider transition towards sustainability.
• Human-centric governance layer: Externally, human-centric governance and organizational vision establish the regulatory and policy framework for implementing the SSLMD. This will include policy alignment, stakeholder engagement, and ethics-based data management. The human-centered approach is designed to guarantee that innovative technologies are compatible with privacy, workers’ well-being, and inclusiveness, by aligning supply-chain innovation with community values and the principles of urban equity.

The framework highlights the feedback loops linking the three dimensions:

• Techno-sustainability feedback loop: Advances in technology help to assess sustainability using real-time performance data (e.g., carbon emissions and energy consumption), and sustainability constraints encourage systems to be cleaner and more efficient. For example, the electric cargo bikes in Amsterdam adjust delivery sequences dynamically to minimize emissions during peak congestion hours [41].
• Techno-human feedback loop: Human contributions (e.g., user feedback and behavioral data) drive AI tuning and design, at the same time as automatization and advanced decision-making tools reinforce user safety, satisfaction, and confidence. For example, in Singapore, real-time feedback from citizens on sidewalks allows for real-time adjustments of drone landing zones or micro-hub shipping schedules [42].
• Governance–sustainability feedback loop: Policy interventions drive sustainable delivering behaviors, which in turn inform sustainability outcomes allowing for adaptive regulations and continuous strategic alignment. For example, Los Angeles has implemented adaptive time slot restrictions for deliveries, using simulation results to reduce traffic congestion while maintaining service levels [43].

For more transparency,

Table 4. A table mapping the elements associated with the SSLMD framework.

Framework Component	Evidence Strength	Type of Support	Studies Ref
Autonomous Vehicles and AI	WE	Empirical, simulation	[41]
IoT and Sensors	WE	Empirical, simulation	[19]
Blockchain	WE	Simulation, conceptual	[19]
Digital Twins	EM	Conceptual, limited situations	[15]
Low-Carbon Logistics	WE	Simulation, conceptual	[10]
Circularity	EM	Simulation, case studies	[44]
Social Equity	SP	Survey-based empirical studies	[29]
Ethics-Based Data Management	SP	Conceptual discussion, conceptual frameworks	[45]
Industry 5.0–Human-Centric Automation	SP	Conceptual discussion, conceptual frameworks	[38]
E-Governance and Digital Participation	EM	Conceptual, limited projects	[24]

indicates the evidence strength for each element. WE (widely evidenced) represents the frequently reported elements in empirical studies. EM (emerging) represents the tested elements in limited simulations. Additionally, SP (speculative) represents conceptual elements with limited empirical support.

4.2. Implementation Conditions

The SSLMD implementation depends on a number of transversal mechanisms:

• Collaborative governing models: the involvement of a variety of stakeholders, including policymakers, drivers, and residents, by means of co-design and living workshops to achieve social acceptance and inclusivity in the system.
• Digital twins and simulation: the use of policy simulation design, where regulations are encoded into digital twin models for testing delivery operational impacts, optimizing routing, and testing regulated actions in virtual worlds before their deployment in the real world. The Singapore case study shows how simulation supports policy adaptation in last-mile urban logistics [46].
• Interoperability and data integration: the harmonizing of information from logistics companies and urban infrastructure to enable transparent coordination.
• Ethical and legal frameworks: the establishment of clear guidance on data sharing, professional transitions, and programmatic transparency to enhance reliability and transparency in intelligent logistics systems.
• AI-based decision-making assistance: the use of machine learning for forecasting demand, preventative maintenance, and adaptive fleet management to improve both environmental performance and efficiency.

4.3. Implications and Outcomes

The proposed SSLMD framework is designed to provide four key systemic benefits:

• Efficiency: reducing operating costs and travel times through AI-based optimization.
• Sustainability: reducing environmental impacts and improving circular logistics practices.
• Equity: Human-centered innovations guaranteeing equitable access, work protection, and integrative governance.
• Resilience: adapting responses to perturbations through data-based anticipation and redundancy.

By combining all these outputs, the framework offers a clear vision for researchers and decision-makers to collaboratively develop the next generation of sustainable logistics systems.

To explicitly illustrate how each aspect of the proposed framework differs according to the different types of cities, as already indicated in the results of the scoping review, namely European cities, American urban areas, megacities in emerging economies, and intermediate cities, we have included brief illustrative examples in

Table A2. Adaptation of SSLMD conceptual framework with city typologies.

City Typology	Framework Adaptation	Guidelines	Illustrative Cases	Ref
European cities	- Governance dimension dominant due to strong municipal authority, LEZs, and enforceable time-window rules. - Technology adoption supported by robust digital infrastructure but constrained by data-protection requirements. - High sustainability integration facilitating micro-hubs, green fleets, and dynamic curb management.	- Prioritize micro-hubs, cargo-bike logistics, and dynamic curb allocation for constrained street networks. - Use strong governance to mandate consolidation schemes and enforce emissions-based fleet differentiation.	Amsterdam, Barcelona: micro-consolidation zones, extensive cargo-bike adoption.	[16,47]
North American urban regions	- Technology dimension central due to reliance on automation (AVs, robots), platform-based optimization. - Governance tools weaker and decentralized, limiting mandatory city-wide policies. - Sustainability varies; often market-driven.	- Focus on autonomous delivery, parcel lockers, ride-pooling logistics, suburban micro-depots. - Emphasize voluntary data sharing and industry partnerships instead of regulation.	Phoenix, Austin: sidewalk robots, drone corridors, suburban micro-depots to handle sprawling geography.	[48]
Emerging cities	- Governance centered on managing informal logistics and gig couriers. - Technology adoption uneven: digital platforms scale rapidly; physical infrastructure lags. - Sustainability focuses on congestion relief and air-quality improvements.	- Deploy scalable, low-cost interventions: shared micro-hubs, informal delivery aggregation, app-based routing optimization.	Delhi, Jakarta: modular micro-hubs; digital routing optimization integrated with informal moto-logistics.	[2]
Intermediate and transition cities	- Balanced adaptation; medium-strength institutions necessitate phased governance development. - Sustainability initiatives often funded externally or piloted by private actors.	- Pilot digital twins, hybrid fleets, and public–private locker networks. - Gradually strengthen governance tools as capacity increases.	Casablanca, Warsaw: hybrid models combining platform-based delivery with emerging environmental regulations.	[49,50]

in Appendix A. This addition clarifies that the framework is not presented as universally uniform, but as an architecture that can be adapted to different contexts, thus directly addressing the reviewer’s concern about generalization.

5. Conclusions and Future Directions

This article presents a scoping review and comprehensive synthesis of the literature on SSLMD systems. By mapping 140 publications systematically and generating their interlinked themes, the review has identified a three-dimensional pattern connecting technological innovation, sustainable integration, and human-centered governance. As a result, the conceptual framework positions SSLMD as a socio-technical ecosystem where smart technologies interplay with green goals, governance mechanisms, and inclusion principles. This interaction allows for participatory design, data-based optimization, and adaptive governance, contributing towards more resilient, efficient, and equitable urban logistics systems. The conclusions emphasize that the search for intelligence must not be pursued independently of the objectives of sustainability and equity.

Despite providing a comprehensive overview of the current state of research, there are several limitations to this study. To start with, it primarily focused on research indexed in Scopus and Wos. Non-English publications and industry studies are excluded, possibly limiting the coverage of applied developments and local best practices. Additionally, the fast advancement of AI, robotics, and data analytics implies that the analyzed data (up to 2025) may become outdated quickly. In addition, this paper focused on developing theoretical insights more than providing empirical support. The proposed SSLMD framework would therefore be subject to testing through modeling applications, piloting, and stakeholder assessment.

On the basis of the gaps identified and the emergent opportunities, the following directions should be prioritized in future research. Primarily, an empirical validation of the SSLMD framework in order to assess its practical viability and adaptability to different types of cities. Furthermore, the integration of multi-agent simulation models or digital twins that incorporate behavioral, economic, and environmental data to illustrate the decision-making processes of operators and decision-makers is needed. Additionally, conducting cross-regional studies to explore how the implementation and acceptance of smart delivery systems are influenced by institutional frameworks, infrastructure readiness, and lifestyle factors is required.