Motion Planning for the ATHLETE Rover with Reinforcement Learning

نویسنده

  • Patrick Mihelich
چکیده

Legged locomotion is attractive because it can enable a robot to traverse far more varied terrain than a wheeled rover is capable of. In the context of planetary exploration, this is especially attractive as the sites of greatest scientific interest tend to be characterized by difficult terrain. For example, it would be extremely difficult for a wheeled rover to make its way into a lunar crater in search of water. Planning legged locomotion is, however, a more difficult problem the wheeled lomotion. Compared to wheeled robots, legged robots tend to have a much larger number of degrees of freedom. A planner for a legged robot therefore has to plan in a high-dimensional configuration space, placing considerable demands on the planner’s efficiency. The problem is further complicated when uneven terrain is considered. Classical motion planning techniques applied to legged locomotion generally assume a flat workspace with obstacles to be avoided. In the case of uneven terrain, there are no explicit obstacles, but a motion planner must take care to maintain the stability of the robot at all times while moving over sloped and uneven surfaces. For articulated legs, contact with the ground creates a closed-loop kinematic chain. Algorithms such as probabilistic roadmaps must be adapted to efficiently handle these closedloop constraints, as a randomly sampled point in the configuration space has zero probability of exactly satisfying them. Especially when the robot is intended for cooperative tasks with humans, an additional problem is planning trajectories which appear natural to a human observer. A planner for a humanoid bipedal robot, for example, might generate bizarre-looking arm motions which aid in balance, even on relatively flat terrain where such motions are unnecessary. A planner for legged locomotion should ideally encode constraints that encourage natural-looking motion. The specific robot considered in this paper is the ATHLETE (All-Terrain Hex-Limbed Extra-Terrestrial Explorer) robot developed by the Jet Propulsion Laboratory (JPL). ATHLETE is intended to be a lunar rover, and is especially designed for movement over broken and uneven terrain. Its hexagonal frame is designed for carrying large Fig. 1. The ATHLETE rover

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A Reinforcement Learning Approach for Motion Planning of Hopping Rovers

This paper considers the problem of navigating a hopping rover on the surface of small Solar System bodies, such as asteroids and comets. Specifically, by controlling the torques applied to three internal flywheels, the rover is able to perform controllable long-range hops and short tumbling maneuvers. The highly stochastic dynamics of bouncing in microgravity and the sequential and discrete na...

متن کامل

Soft Value Iteration Networks for Planetary Rover Path Planning

Value iteration networks are an approximation of the value iteration (VI) algorithm implemented with convolutional neural networks to make VI fully differentiable. In this work, we study these networks in the context of robot motion planning, with a focus on applications to planetary rovers. The key challenging task in learningbased motion planning is to learn a transformation from terrain obse...

متن کامل

Stochastic Motion Planning for Hopping Rovers on Small Solar System Bodies

Hopping rovers have emerged as a promising platform for the future surface exploration of small Solar System bodies, such as asteroids and comets. However, hopping dynamics are governed by nonlinear gravity fields and stochastic bouncing on highly irregular surfaces, which pose several challenges for traditional motion planning methods. This paper presents the first ever discussion of motion pl...

متن کامل

FootFall: A Ground Based Operations Toolset Enabling Walking for the ATHLETE Rover

The ATHLETE (All-Terrain Hex-Limbed Extra-Terrestrial Explorer) vehicle consists of six identical, six degree of freedom limbs. FootFall is a ground tool for ATHLETE intended to provide an operator with integrated situational awareness, terrain reconstruction, stability and safety analysis, motion planning, and decision support capabilities to enable the efficient generation of flight software ...

متن کامل

Software tools for the cognitive development of autonomous robots

Knowledge representation and reasoning 4 Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Probabilistic formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Fuzzy logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Planning 6 Motion planning . . . . . . ...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2007