An Efficient Locally Reactive Controller for Safe Navigation in Visual Teach and Repeat Missions

Along Similar Lines: Local Obstacle Avoidance for Long-Term Autonomous Path Following

“…Mattamala et al [11] explore the use of a locally reactive controller for completing visual teach-and-repeat missions in the presence of obstacles. Their research employs local elevation maps to compute vector representations of the environment and directly generate control twist commands using a Riemannian Motion Policies controller.…”

Fast Traversability Estimation for Wild Visual Navigation

“…A median filter removed undesired noise and artifacts before the distance transform method [10] was used to obtain a Signed Distance Field (SDF), which represents the distance to the closest untraversable object. We implemented a local planner method based on Mattamala et al [23], which exploits the SDF and the local goal to generate a SE(2) twist command which drives the robot towards the goal while avoiding untraversable terrain. Finally, the twist command becomes the input to a robust learning-based locomotion controller based on the work by Miki et al [25], which is able to traverse rough terrain typically inaccessible to wheeled robots.…”

Learning Forward Dynamics Model and Informed Trajectory Sampler for Safe Quadruped Navigation

“…Post-processing the planned path heuristically [50] for smoothness still cannot be a complete solution to guarantee safety in local planning. Some works used reactive methods [37,30,18] with vector field representation for local obstacle avoidance, but they result in jerky and discrete command changes due to the lack of long horizon planning. Gilroy et al [11] and Li et al [35] used collocationbased trajectory optimization to handle long horizon planning, but the decoupled nature between the local planner and the command-tracking controller can still cause a collision in complex environments because the local planner cannot take into account the performance characteristics of the commandtracking controller.…”