Robotics & Autonomous Vehicles
Perception that keeps pace with autonomy
Prophesee’s Event-Based Vision provides low-latency, data-efficient perception for robots, civil drones and autonomous vehicles operating within tight compute, bandwidth and power budgets.
Navigate, localize and react to changing surroundings in real time – even during rapid motion and across difficult lighting.
Performance for responsive autonomy
Capture rapid motion and sudden changes continuously
>10 kHz
equivalent temporal precision
Maintain perception through difficult lighting and sudden transitions
>120 dB
dynamic range
Give autonomous systems more time to interpret and respond
<1 ms
latency
Enable ultra-low-power, always-on perception
<2 mW
sensing power
Robotics and autonomous-vehicle applications
Autonomous navigation and obstacle avoidance
Prophesee’s Event-Based Vision helps robots and autonomous vehicles navigate changing environments and avoid obstacles in real time. Precisely timed information about movement and change supports responsive closed-loop control, even during rapid platform motion or in unfamiliar surroundings. Sparse output also limits the onboard processing required for continuous perception.
Typical use cases
- Autonomous navigation
- Reactive obstacle avoidance
- AGV and AMR navigation
- Civil drone navigation
- Mobile robot safety
- Autonomous flight in unknown environments
Autonomous drone obstacle avoidance
Navigate around static obstacles using an onboard monocular event camera and a reactive events-to-control policy demonstrated in indoor and outdoor flight by UZH Robotics and Perception Group. Learn more
Visual odometry, SLAM and localization
Maintain accurate motion and position estimates through rapid maneuvers and GNSS-denied operation. Event-based sensing preserves fine temporal detail for visual odometry, sensor fusion and localization, helping autonomous systems track their state continuously as they move.
Typical use cases
- Visual odometry
- Visual-inertial odometry
- Simultaneous localization and mapping
- Robot localization
- Pose and motion estimation
- GNSS-denied navigation
High-speed object detection and tracking
Detect and track moving objects continuously, with the temporal detail needed to estimate trajectories and respond sooner. Event-Based Vision follows changes in position and direction as they unfold, helping autonomous systems maintain awareness of dynamic objects throughout their movement.
Typical use cases
- Dynamic object detection
- Continuous object tracking
- Trajectory estimation
- Collision prediction
- Pedestrian and cyclist detection
- Vehicle tracking
- Human and robot tracking
Detection and tracking in dynamic scenes
Detect and track moving road users and objects continuously, including across scenes with difficult contrast and changing illumination.
Autonomous driving and ADAS
Autonomous-driving and driver-assistance systems must interpret rapidly changing road environments with enough time to support braking, avoidance and other safety responses.
Event-Based Vision can complement RGB cameras, radar, lidar and other sensing modalities with precise motion and temporal information – supporting low-latency near-field perception without adding a constant full-frame data stream load.
Typical use cases
- Autonomous driving
- Advanced driver-assistance systems
- Emergency braking
- Near-field collision avoidance
- Vulnerable-road-user detection
- Urban hazard detection
Emergency braking
Detect rapidly developing road hazards and provide timely motion information for autonomous-driving and driver-assistance functions.
Perception across changing conditions
Responsive autonomy depends on reliable motion information across the operating environment. See how Event-Based Vision performs through difficult lighting and captures temporal signals that conventional frame rates may not resolve.
Related automotive application: In-cabin sensing
Event-Based Vision can capture subtle driver and occupant movement across changing cabin lighting, supporting driver monitoring, occupant classification and attention analysis.
Event-driven robotic vision and touch
Combining precise visual motion information with neuromorphic touch can help robots identify objects, detect contact and adjust their grip during manipulation.
Research from the National University of Singapore demonstrates how visual-tactile sensing can support object recognition and grasping, including rotational slip detection in 0.08 seconds.
Typical use cases
- Object recognition and grasping
- Grip control
- Slip detection
- Dexterous manipulation
- Visual-tactile learning
Building fixed production automation or high-speed pick-and-place systems?
Developing perception for a defense platform or counter-UAS mission?
Build autonomous perception
Evaluate Event-Based Vision, develop application intelligence and integrate responsive perception into embedded robotics and mobility systems.
Embra®
An Event-Based Vision Exploration kit featuring a compact, industrial grade USB camera for capturing application data, evaluating performance and developing robotics and autonomous-system prototypes.
Hearth®
Prophesee’s next-generation software platform for developing, integrating and deploying Event-Based Vision applications on embedded systems.
Integration support
Work with Prophesee experts to adapt Event-Based Vision to your platform, autonomy stack, operating environment and deployment requirements.
Bring faster perception to your autonomous system
Prophesee experts can help assess your autonomy requirements and define a path from feasibility testing to embedded integration and deployment.