For many people, LiDAR is still strongly associated with autonomous vehicles: cars scanning roads, detecting pedestrians, cyclists, lane boundaries, and obstacles. This association makes sense because autonomous driving has been one of the most visible areas for LiDAR development. However, the future of LiDAR is not limited to vehicles.
A more interesting question is what happens when intelligence is also built into the infrastructure around the vehicle. Cities could place LiDAR sensors on traffic lights, poles, roadside units, buildings, campuses, and public spaces. From these fixed and elevated positions, LiDAR can help cities understand how pedestrians, vehicles, and cyclists move through shared environments. This is where LiDAR for smart cities becomes powerful: it can move urban infrastructure beyond simple detection toward motion understanding, safety analysis, and privacy-aware public-space monitoring.
This post was written by MSc Nawfal Guefrachi, a Ph.D. student at Missouri S&T researching LiDAR-based sensing, smart-city perception, and AI-driven activity understanding.
Table of Contents
Why LiDAR for Smart Cities?
A city is not a static map. Pedestrians cross intersections, vehicles turn through traffic, cyclists move between lanes, and crowds form near campuses, transit stations, and downtown areas. A smart city therefore needs more than vehicle counts or congestion maps. It needs to understand what is happening in real time.
Consider a pedestrian crossing while a vehicle prepares to turn. A traditional traffic signal follows a fixed timing plan. An intelligent intersection could ask better questions: Is someone still in the crosswalk? Is a vehicle moving toward a conflict zone? Is a pedestrian moving unusually slowly or suddenly changing direction? A LiDAR sensor mounted on a traffic light or street pole can observe the scene from a stable viewpoint and measure 3D position, distance, and motion. LiDAR and intelligent transportation technologies have already been discussed as promising tools for smart-city monitoring, traffic surveillance, and pedestrian safety applications [1], [2].
Why Elevated LiDAR Matters?
If LiDAR for smart cities is about helping infrastructure understand movement, sensor placement becomes critical. Vehicle-mounted LiDAR only sees the world from the vehicle’s position, and its view can be blocked by buses, parked cars, pedestrians, trees, or roadside objects. An elevated LiDAR sensor mounted on a traffic light, pole, roadside unit, or building can observe the same area from above and provide a more stable view.
Infrastructure Viewpoint and Continuous Monitoring
This elevated perspective is especially useful at intersections. From a vehicle’s viewpoint, parts of the scene may be blocked or visible only briefly. From an elevated infrastructure viewpoint, the system can observe pedestrians, vehicles, and cyclists more globally and understand how they move relative to one another.
Figure 1 illustrates this concept. Instead of sensing the street only from a vehicle perspective, LiDAR is mounted on urban infrastructure and observes pedestrians, vehicles, and activities from an elevated viewpoint.
Because the sensor is fixed, it can monitor the same crosswalk, sidewalk, campus walkway, or transit area over time. This supports real-time detection and long-term safety analysis. Elevated and roadside LiDAR systems have already been explored for pedestrian and vehicle detection, tracking, and infrastructure-based perception in intelligent transportation systems [3], [4].
Placement Tradeoffs and Point Cloud Quality
Elevation is useful, but it is not automatically ideal. A sensor mounted too low may suffer from occlusion, while one mounted too high may cover a wider area but produce sparser pedestrian point clouds. A crosswalk-monitoring sensor may need enough point density to capture posture, while a traffic-flow sensor may prioritize wide coverage. Height, angle, field of view, and scanning resolution therefore determine what the system can actually understand.
Seeing Motion in Three Dimensions
LiDAR provides 3D point clouds, where each point represents a measured location in space. Cameras capture appearance, color, texture, and visual context. LiDAR captures geometry, distance, and spatial structure. For pedestrian monitoring, this matters because a pedestrian is not only a shape in an image; they are a moving 3D object with position, height, posture, speed, and trajectory.
Many safety questions are motion questions. Is someone walking normally, stopping suddenly, or falling in a public space? With several LiDAR frames, an AI system can analyze changes in position, posture, velocity, and trajectory. LiDAR can also provide reflectivity, or intensity, which describes how strongly a surface returns the laser signal. Reflectivity should not replace geometry or motion, but it can add useful cues in complex urban scenes.
Privacy-Aware Urban Perception
Smart-city sensing also raises a human question: how can cities improve safety without making people feel constantly watched? This is one of the strongest arguments for LiDAR for smart cities. Cameras are powerful, but in public spaces they may capture identifiable details such as faces, clothing, and personal appearance.
LiDAR offers a different representation. It captures sparse geometric point clouds rather than detailed visual images. A LiDAR system may still detect and track pedestrians, estimate motion, and analyze activity, but without relying on facial or appearance information. This does not mean LiDAR automatically solves every privacy concern. Public sensing still requires careful policies for storage, access, transparency, and ethical deployment. However, LiDAR can help shift the conversation from identity-based surveillance toward behavior-aware sensing [2], [5], [6].
Cameras, LiDAR, or Both?
If camera-based perception systems have become so powerful, why should cities still consider LiDAR? The answer is not “camera versus LiDAR,” but rather what type of information a city needs. Cameras are highly effective for visual context: they can recognize traffic light colors, road signs, vehicle shapes, pedestrian appearance, and scene details. Modern camera-based systems have also shown strong performance in pedestrian detection, tracking, and activity recognition [7].
However, public-space monitoring also requires reliable 3D motion understanding. Camera systems can estimate depth and motion, but their performance may be affected by lighting changes, viewpoint variation, occlusion, and depth uncertainty [8], [9]. LiDAR also faces challenges such as occlusion and point sparsity, but it directly measures 3D structure. A pedestrian is represented less by appearance and more by position, distance, direction, speed, and posture. Prior work has shown the value of LiDAR for pedestrian detection, tracking, gait analysis, and human activity recognition [3], [10]–[13].
LiDAR also offers an important privacy advantage in public environments. Since it captures geometric point clouds rather than detailed facial or clothing appearance, it can support pedestrian monitoring while reducing the amount of identifiable visual information collected. Therefore, cameras remain valuable when visual context is needed, while LiDAR is especially useful for 3D motion, distance-aware sensing, varied lighting conditions, and privacy-aware monitoring. In many smart-city settings, the strongest solution may combine both.
From Detection to Motion Understanding
Whether a system uses LiDAR alone or combines it with cameras, the deeper goal is the same: moving from detection to understanding. Detection asks what objects are in the scene. Motion understanding asks what those objects are doing.
For pedestrian safety, this distinction matters. A person standing near a sidewalk, crossing an intersection, running suddenly, or falling may all appear as pedestrians, but each situation has a different meaning. LiDAR point clouds can support this deeper understanding because they provide spatial and temporal information. Across several frames, the system can analyze position, posture, velocity, trajectory, and movement patterns.
This is the direction of my own research on elevated LiDAR-based pedestrian activity monitoring, where 3D point clouds from an elevated viewpoint are used to detect pedestrians, track their motion, and classify activity patterns [15]. Figure 3 shows this idea as a sensing-to-decision pipeline.
This pipeline can support public-space fall detection, campus safety, smart intersection monitoring, and pedestrian behavior analysis. Recent LiDAR-based pedestrian monitoring work shows how infrastructure-mounted sensors can support both detection and behavior understanding [2].
Training AI on Real LiDAR Data
For AI to understand pedestrian motion, it first needs examples. That sounds simple, but in practice it is one of the biggest challenges for LiDAR for smart cities. A city may naturally capture many examples of normal walking, crossing, or waiting. But what about the rare events that matter most for safety? How many examples of falling, limping, sudden abnormal motion, or near-miss events can be collected safely, ethically, and at scale?
This is where the data problem begins. AI models learn from patterns, and if the most important events are rare, the model may not see enough of them during training. Real LiDAR data is also difficult to label frame by frame, especially when the goal is not only to detect a pedestrian, but to understand motion, posture, and activity over time.
Simulation helps address part of this challenge. Researchers can design scenes, control pedestrian behavior, adjust sensor placement, and generate LiDAR-like point clouds for specific activities. Simulators such as Blender, CARLA, Webots, Gazebo, AirSim, and NVIDIA Isaac Sim can support this process. Blender, for example, can create customized urban scenes, animate pedestrians, and generate LiDAR-like data through raycasting or LiDAR simulation tools [16].
Simulation is useful, but it is not the final answer. Real LiDAR data includes noise, sparsity, reflectivity variation, weather effects, calibration errors, occlusion, and unpredictable behavior. A simulated pedestrian may move cleanly, while a real person may stop suddenly, carry bags, walk in groups, or be partially hidden. The practical approach is to learn structure in simulation, then adapt and validate with real-world data.
Deployment also raises practical questions. Where should LiDAR be mounted? How high is too high? Who maintains the sensors? How is the data processed? Will the public trust the system? For LiDAR for smart cities to move from research to deployment, the system must be reliable, affordable to maintain, useful for city operations, and acceptable to the people who move through those spaces every day.
Applications of LiDAR for Smart Cities
After considering sensing, privacy, fusion, AI, and deployment, where can LiDAR for smart cities make a visible difference? At a smart intersection, LiDAR could monitor pedestrians, vehicles, and cyclists in 3D. If a pedestrian is still in the crosswalk when the light is about to change, signal timing could be extended. If a vehicle and pedestrian move toward the same conflict zone, the system could trigger a warning or log the event for safety analysis. Roadside and infrastructure-based LiDAR studies have shown the potential of LiDAR for detection and tracking in traffic environments [3], [4].
In campuses, sidewalks, transit stations, and parking areas, LiDAR could help detect unusual pedestrian motion or falls without relying on identifiable camera footage. Instead of recording facial appearance, the system can focus on geometric movement patterns such as standing, walking, running, sitting, or falling. Human activity recognition and gait-related LiDAR studies suggest that geometric point cloud data can support human motion analysis [5], [6], [13], [12].
LiDAR can also support crowd and mobility analysis. At a football game, concert, graduation ceremony, or subway station, the key question is not only how many people are present, but how they are moving. LiDAR-based mobility analysis could help planners design public spaces that feel safer, smoother, and more comfortable.
From Concept to Real-World Deployment
Real-world deployments suggest that LiDAR-based smart-city monitoring is moving from concept toward practice, but many applications are still in an early deployment or research-driven stage. The main promise is clear: fixed LiDAR sensors can help cities measure road-user position, distance, trajectory, speed, and movement patterns in 3D, which is useful for intersection safety, traffic management, and pedestrian monitoring. Several current examples show this direction. In Peachtree Corners, Georgia, Curiosity Lab and Opsys have introduced solid-state LiDAR technology in a real smart-city test environment [17]. Ouster’s BlueCity platform provides LiDAR-based traffic detection and safety analytics for smart infrastructure [18], and Ouster is also involved in a Chattanooga smart-corridor deployment that integrates digital LiDAR sensors and edge AI after an initial pilot at 12 intersections [19]. Other examples include a LiDAR-powered smart intersection in Utah using Seoul Robotics technology [20], a Helsinki traffic-safety project using Velodyne’s Intelligent Infrastructure Solution for detecting vehicles, pedestrians, and cyclists [21], and a Busan smart-city project using Quanergy’s 3D LiDAR solution for road and pedestrian safety monitoring [22].
At the same time, these examples also show that the field is still developing. Many deployments focus on traffic flow, object detection, safety-event analytics, or digital-twin-style intersection monitoring, while more advanced tasks such as robust pedestrian activity recognition, abnormal behavior detection, and long-term human motion understanding still require further validation. Broader adoption will depend on lower deployment costs, easier integration with existing traffic infrastructure, stronger real-world testing across different cities and weather conditions, and more reliable AI models. This is why training data and simulation remain important. Real-world LiDAR data for pedestrian activity analysis can be difficult to collect, label, and generalize, so realistic simulation can help generate diverse scenarios, test edge cases, and improve model robustness before large-scale deployment.
Conclusion
The larger story is that LiDAR is no longer only about making vehicles smarter. It may also help make infrastructure more aware. LiDAR for smart cities can support pedestrian safety, traffic analysis, emergency response, privacy-aware monitoring, and a deeper understanding of public-space motion.
The question is not whether LiDAR should replace cameras. The question is how each sensor can contribute to safer and more intelligent public spaces. Cameras provide rich visual context. LiDAR provides direct 3D geometry, motion information, and privacy-aware spatial sensing. Together, or in carefully selected applications, these technologies can help cities move beyond simple detection toward understanding what is happening.
About the Author
Nawfal Guefrachi is a Ph.D. student whose research focuses on elevated LiDAR sensing, smart-city perception, pedestrian monitoring, and AI-driven human activity understanding. His work explores how 3D LiDAR data can support traffic monitoring, pedestrian safety, abnormal activity detection, simulated data generation, and deep learning-based perception systems.
His related publications include several IEEE conference papers, as well as more recently accepted journal work in prestigious peer-reviewed venues. Readers interested in learning more about his research can explore his publications on LiDAR-based perception, smart-city sensing, and AI-driven activity recognition.
Google Scholar Profile: Nawfal Guefrachi – Google Scholar
References
[1] I. Ahmed, G. Jeon, and A. Chehri, “A Smart IoT Enabled End-to-End 3D Object Detection System for Autonomous Vehicles,” IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 11, pp. 13078–13087, Oct. 2022.
[2] O. Rinchi, H. Ghazzai, A. Alsharoa, and Y. Massoud, “LiDAR Technology for Human Activity Recognition: Outlooks and Challenges,” IEEE Internet of Things Magazine, vol. 6, no. 2, pp. 143–150, 2023.
[3] J. Zhao, H. Xu, H. Liu, J. Wu, Y. Zheng, and D. Wu, “Detection and Tracking of Pedestrians and Vehicles Using Roadside LiDAR Sensors,” Transportation Research Part C: Emerging Technologies, vol. 100, pp. 68–87, 2019.
[4] H. Yu, Y. Luo, M. Shu, Y. Huo, Z. Yang, Y. Shi, et al., “DAIR-V2X: A Large-Scale Dataset for Vehicle-Infrastructure Cooperative 3D Object Detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 21329–21338.
[5] M. A. U. Alam, F. Mazzoni, M. M. Rahman, and J. Q. Widberg, “LAMAR: LiDAR Based Multi-Inhabitant Activity Recognition,” in Proceedings of MobiQuitous 2020: 17th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services, 2020, pp. 1–9.
[6] M. A. U. Alam, F. Mazzoni, M. M. Rahman, and J. Q. Widberg, “Towards Adaptive Multi-Inhabitant Activity Recognition in Smart Homes,” in IEEE INFOCOM 2021 – IEEE Conference on Computer Communications Workshops, 2021.
[7] U. Gawande, K. Hajari, and Y. Golhar, “Real-Time Deep Learning Approach for Pedestrian Detection and Suspicious Activity Recognition,” Procedia Computer Science, vol. 218, pp. 2438–2447, 2023.
[8] D. Zhang, Y. Shao, Y. Mei, H. Chu, X. Zhang, H. Zhan, and Y. Rao, “Using YOLO-Based Pedestrian Detection for Monitoring UAV,” in Tenth International Conference on Graphics and Image Processing (ICGIP 2018), Proc. SPIE 11069, 2019.
[9] T. Ophoff, K. Van Beeck, and T. Goedemé, “Improving Real-Time Pedestrian Detectors with RGB+Depth Fusion,” in Proceedings of the IEEE International Conference on Image Processing Theory, Tools and Applications, 2018.
[10] C. Benedek, B. Gálai, B. Nagy, and Z. Jankó, “Lidar-Based Gait Analysis and Activity Recognition in a 4D Surveillance System,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 28, no. 1, pp. 101–113, 2018.
[11] W. Wang, X. Chang, J. Yang, and G. Xu, “LiDAR-Based Dense Pedestrian Detection and Tracking,” Applied Sciences, vol. 12, no. 4, Article 1799, 2022.
[12] C. Shen, C. Fan, W. Wu, R. Wang, G. Q. Huang, and S. Yu, “LidarGait: Benchmarking 3D Gait Recognition with Point Clouds,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023, pp. 1054–1063.
[13] H. Yamada, J. Ahn, O. M. Mozos, Y. Iwashita, and R. Kurazume, “Gait-Based Person Identification Using 3D LiDAR and Long Short-Term Memory Deep Networks,” Advanced Robotics, vol. 34, no. 18, pp. 1201–1211, 2020.
[14] O. H. Jafari and M. Y. Yang, “Real-Time RGB-D Based Template Matching Pedestrian Detection,” in Proceedings of the IEEE International Conference on Robotics and Automation Workshops, 2014.
[15] N. Guefrachi, H. Ghazzai, A. Alsharoa, and G. Setti, “Leveraging 3D LiDAR Sensors to Enable Enhanced Urban Safety and Public Health: Pedestrian Monitoring and Abnormal Activity Detection,” 2024.
[16] A. Reitmann, M. Neumann, and B. Jung, “Blainder: A Blender LiDAR Simulator for Computer Vision Applications,” 2021.
[17] Curiosity Lab at Peachtree Corners, “Opsys Collaborates with Curiosity Lab and City of Peachtree Corners to Launch Solid-State LiDAR Technology,” Nov. 2024. [Online]. Available: https://curiositylabptc.com/press_release/opsys-collaborates-with-curiosity-lab-and-city-of-peachtree-corners-to-launch-solid-state-lidar-technology/
[18] Ouster, “Ouster BlueCity: AI-Driven Lidar for Smarter Cities & Traffic Management.” [Online]. Available: https://ouster.com/products/software/bluecity
[19] Ouster, “ITS International Chattanooga Case Study: Smart Corridor Lidar Traffic Installation.” [Online]. Available: https://ouster.com/insights/case-studies/its-chattanooga-lidar-traffic-installation
[20] SmartCitiesWorld, “Utah DoT Installs Lidar Tech at Salt Lake City Intersection,” Jun. 2024. [Online]. Available: https://www.smartcitiesworld.net/digital-twins/utah-dot-installs-lidar-tech-at-salt-lake-city-intersection
[21] Velodyne Lidar, “Velodyne Lidar’s Intelligent Infrastructure Solution Deployed in Helsinki Traffic Safety Improvement Project,” Jun. 2022. [Online]. Available: https://via.tt.se/pressmeddelande/3325637/velodyne-lidars-intelligent-infrastructure-solution-deployed-in-helsinki-traffic-safety-improvement-project
[22] Quanergy, “Busan Smart City Project Paves the Way for Safer Roads with LiDAR-Based IoT Solution,” Oct. 2024. [Online]. Available: https://quanergy.com/busan-smart-city-project-paves-the-way-for-safer-roads-with-lidar-based-iot-solution/
