Based on 3D time-of-flight technology, CoPilot constructs a local map of its environment in real-time, estimates the local ground plane, and segments objects of interest (e.g., walls, curbs, obstacles) from the background. Beyond geometric object segmentation, CoPilot also employs a real-time terrain classification system used to identify driveable surfaces.
CoPilot employs a sample-based path planner. The planner samples from a feasible set of linear and angular velocities achievable by the vehicle over a configurable time horizon. The sampled trajectories are evaluated against a cost function to select an optimal path of travel while ensuring safety in the presence of obstacles and drop-offs.
CoPilot integrates deeply with the vehicle’s motor controller providing drive-by-wire capabilities to the system. Sensor data are fused to estimate the vehicle’s instantaneous linear and angular velocities. Based on these data, a PID controller ensures responsiveness and accurate regulation of the vehicle’s velocity.
Love Park Robotics is actively seeking OEM partners to bring our technology to market. We would be happy to discuss the specifics of your application to evaluate the suitability of CoPilot. Target applications for CoPilot include, but are not limited to, active driving aids for electric-powered wheelchairs, automating service delivery in hospitals or other institutions (e.g., meds, linens, food), adding semi-autonomous control to tele-presence robots operating over high latency networks (e.g., the Internet), and other intelligent consumer products (indoor or outdoor).