Advance Warning of Vehicle–Pedestrian Conflicts at Right Turns and Crosswalks
Solutions
Lookout™
Industry
Roads·Rail
Applicatios
Collision & Pedestrian Safety
Turn intersection safety, once reliant on drivers’ attention,
into a system that infrastructure protects at all times.
Target Customers
• Local government traffic and road safety departments
• Road corporations and transportation facility operators
• Intelligent transportation system (ITS) and smart crosswalk implementation providers
• Traffic safety and accident analysis specialist organizations
Challenges
• Vehicle–pedestrian conflicts recur at right turns and unsignalized crosswalks, yet there is no continuous monitoring means to prevent accidents in advance
• Vehicle-mounted sensors perceive only their own driving path and miss pedestrians and two-wheelers in intersection blind spots
• Accident prevention ultimately depends on individual drivers’ attention, so risk rises sharply at night and in bad weather
Existing Limitations
• Signals, signs, and road markings are passive means that depend on drivers’ perception and cannot fundamentally eliminate blind spots
• CCTV is vulnerable to nighttime, backlight, and occlusion and has no real-time risk assessment capability
• Vehicle-only sensors make it difficult to comprehensively assess intersection conflict situations from an infrastructure perspective
Solution Approach
• Detect vehicles and pedestrians approaching the intersection in 3D with LiDAR and track their trajectories
• Assess collision risk from distance, speed, and trajectory and issue advance warnings by TTC-based risk level
• Define hazard areas such as right-turn blind spots and unsignalized crosswalks as virtual zones for continuous monitoring
• Integrate with existing safety facilities such as traffic signals, warning lights, and pedestrian alerts
• Process 3D coordinates rather than raw video to minimize privacy exposure
Expected Benefits
• Prevent vehicle–pedestrian conflict accidents in blind spots
• Break away from reliance on driver attention through continuous infrastructure monitoring
• Maintain the same detection performance at night and in bad weather
• Secure a basis for improving accident-prone locations through risk event records
