CAMEL: Learning Cost-maps Made Easy for Off-road Driving
This addresses the problem of robust and adaptive path planning for robotic vehicles in off-road settings, representing an incremental improvement over handcrafted methods.
The paper tackles the challenge of manually tuning cost-maps for robotic path planning in off-road environments by proposing CAMEL, a deep learning framework that learns cost-map values from demonstrations, resulting in collision-free vehicle motion in unstructured terrains as validated in simulations and real-world tests.
Cost-maps are used by robotic vehicles to plan collision-free paths. The cost associated with each cell in the map represents the sensed environment information which is often determined manually after several trial-and-error efforts. In off-road environments, due to the presence of several types of features, it is challenging to handcraft the cost values associated with each feature. Moreover, different handcrafted cost values can lead to different paths for the same environment which is not desirable. In this paper, we address the problem of learning the cost-map values from the sensed environment for robust vehicle path planning. We propose a novel framework called as CAMEL using deep learning approach that learns the parameters through demonstrations yielding an adaptive and robust cost-map for path planning. CAMEL has been trained on multi-modal datasets such as RELLIS-3D. The evaluation of CAMEL is carried out on an off-road scene simulator (MAVS) and on field data from IISER-B campus. We also perform realworld implementation of CAMEL on a ground rover. The results shows flexible and robust motion of the vehicle without collisions in unstructured terrains.