Infrastructure-Aided Localization and State Estimation for Autonomous Mobile Robots

A slip-aware localization framework is proposed for mobile robots experiencing wheel slip in dynamic environments. The fuses infrastructure-aided visual tracking data (via fisheye lenses) and proprioceptive sensory from a skid-steer robot to enhance accuracy reduce variance of the estimated states. includes: thread detect track stereo image through computationally efficient 3D point cloud generation using region interest; ego motion which uses odometry mechanism estimate pose utilizing model considering slip. Covariance intersection used fuse prediction (using data) thread, such that updated remains consistent. As confirmed by experiments on robot, designed addresses state estimation challenges indoor/outdoor autonomous experience high-slip, uneven torque distribution at each (by planner), or occlusion when observed an infrastructure-mounted camera. system real-time capable scalable multiple environmental cameras.