Planning and Control for Autonomous Drives of the Mars Sample Recovery Helicopter

The planned Mars Sample Return (MSR) campaign is humanity’s first attempt at returning samples from the Martian surface to Earth. This campaign is composed of multiple Martian surface assets, including the Perseverance rover and the Sample Retrieval Lander; which collaborate to collect, retrieve, and ultimately return a diverse set of Martian rock and regolith samples. The Sample Recovery Helicopter (SRH) is a backup system of two hybrid aerial-ground robots that transport sample tubes from the Three Forks sample depot to the Sample Retrieval Lander (SRL) in the case that the Perservance rover is not able to transport its own cache of samples to SRL.An SRH robot is a hybrid aerial-ground vehicle, capable of flying between Three Forks and SRL, performing accurate drives from landing sites to tube pickup and dropoff locations, and handling sample tubes with a manipulator arm. With the weight constraint induced from flying in the Martian atmosphere, the ground mobility and sample manipulation subsystems are each subject to significant mass constraints (< 150g each). The mobility system has a four wheel skid-steer configuration, with 10 cm diameter wheels, and a maximum rock height traversal capability of 3.5 cm. To enable the manipulator arm to pickup the tube, the robot must drive to within a pose error of ±2 cm in x, ±4cm in y, and ±5◦ in yaw. With this tight accuracy requirement, the low control authority of lightweight skid-steer vehicles, and low rock-height traversal, the mobility system requires a highly capable autonomous driving capability.This paper presents the design of the planning and control strategies for SRH autonomous driving, along with a quantitative evaluation of the platform’s ability to drive to a desired pose on Martian analogue terrain. The ABIT* anytime optimizing sampling-based planner is used to plan feasible and distance-efficient paths assuming an occupancy grid representation of the local environment. A path-following controller is then used to follow the generated path by using spot turn and straight drive motion primitives. A separate controller is used for the final approach to a target pose using drives along closed-loop shallow arcs. Performance evaluation results show that such a strategy reliably drives the skid-steer platform to within the required pose error.The decision to implement Mars Sample Return will not be finalized until NASA’s completion of the National Environmental Policy Act (NEPA) process. This document is being made available for information purposes only.