Research On Path Planning Technology For Lunar Rover Based On The Human-Machine Cooperation

Navigation and control of Lunar Rover (LR) is one of the key technologies in the development of LR, which is an important part in lunar exploration project. Path planning is one of the key problems for safety LR navigation on the uncertain and unstructured lunar environment. Besides the length of the path, other factors including the traversability of the terrain, the uncertainty of obstacles, should be considered in LR path planning. In consideration of the significance of human-machine intelligence integration for the performance improvement of LR navigation, the human-machine cooperation based path planning method is studied as the emphasis of the thesis, which is discussed as a multi-objective optimization problem under multiple constrains.The main contributions of this thesis are as follows:First, based on human-machine cooperation concept, a three-layer hierarchical architecture for LR navigation and control system is proposed. In the architecture, a deliberation-reaction hybrid navigation and control strategy is adopted. Various human-machine coordination mechanisms are developed, so the competence of human operator on both planning and reaction are incorporated into the ground-based cooperation planning layer, and imported into the onboard navigation system through tele-control command , thus the human-machine intelligence integration is implemented, the movement safety and the environment adaptability of lunar rover are improved.Second, an environment modeling method is proposed, which is aimed at the multi-constrained path planning based on the human-machine cooperation. Four factors are taken into consideration in the modeling method: the traversability is analyzed with lunar Digital Elevation Model (DEM), together with the constraints of traveling performance of the rover; the uncertainty of environment perception and controller execution are modeled into the probability distribution model of the obstacles, which is established using the concept of latent risky area and risk measurement; as for the human-machine interaction in the process of path planning, human intelligence is integrated into the environment model by introducing the path guidance point and the path guiding function; the movement stability of LR is guaranteed by modeling the measurement of path smoothness.Third, a path evaluation method is proposed for the Multi-Objective Path Planning(MOPP) based on human-machine cooperation. Then a MOPP optimization model is established on the basis of the multi-objective optimization theory. Further more, a Multi-Objective Path Planning Evolution Algorithm (MOPPEA) is put forward, which can provide many non-dominated solutions at the same time.Fourth, a path-planning parameter selection method is proposed based on the fuzzy preference relation, and the multi-objective decision problem is transformed into a single objective decision problem in accordance with human intention. Using the complete consistency constraints, a fast algorithm to construct the fuzzy preference relation matrix is developed. This algorithm not only approximates effectively the operator's subjective preference, but also reduces the blindness in evaluating among various optimizing objectives.Based on the work above, a path planning system based on human-machine cooperation is designed and implemented, and the performance is tested using a prototype lunar rover in outdoor, natural terrain environment. The experiment results verify the feasibility and the validity of the methods discussed above.