越野环境下存在着多种类型、状态和形态的障碍物、地形、运动车辆和环境威胁等影响车辆运动安全的环境要素，在复杂的越野环境中规划安全、高效、可行的路径和行车轨迹是无人地面车辆的关键技术。传统的车辆运动规划方法缺乏对多种复杂越野地形、环境威胁和运动车辆风险的有效评估方法，运动规划过程中，难以权衡路径的安全性与通行效率，无法解决规划时间与轨迹优化之间的矛盾，不能适应越野环境下的运动规划要求。针对这些问题，本文建立了以越野条件下环境风险评估为基础的车辆运动规划架构，提出了基于场理论的统一化越野环境态势场风险量化评估方法。基于环境态势场模型，提出了多目标均衡式概率图全局路径规划和搜索树渐进优化式局部轨迹规划方法，能够在复杂越野环境下规划安全、高效、可行的路径和行车轨迹。首先，提出了考虑越野地形与环境威胁等多种要素的环境态势场模型。针对环境边界、静态障碍物、越野地形、运动车辆和环境威胁等要素的种类、形态、状态各异，难以量化评估的问题，按其各自的特点和变化规律，建立统一化环境态势场模型，用场的强度、作用范围和方向等特征来量化评估多个要素在越野环境中的作用，实现了越野环境下的车辆运动风险统一量化评估。其次，提出了基于环境态势场的多目标均衡式概率图全局路径规划方法。用概率采样方法生成空间拓扑图，通过评估拓扑图节点连接的可行性、距离和环境态势，建立多维度通行评估矩阵。用评估矩阵分析和计算车辆运动扩展路径的通行成本，并结合车辆自身的越野、防护性能和目标任务需求，均衡越野环境中车辆通行效率与安全的影响，实现了异质车辆多目标均衡式全局路径规划。然后，提出了实时风险评估的搜索树渐进优化式局部轨迹规划方法。用增量式采样方法生成空间随机搜索树，基于车辆的运动学约束条件扩展搜索树轨迹，用环境态势场模型评估车辆在运动轨迹上的行驶风险，用运动轨迹长度和转角评估运动效率。通过动态轨迹重构与动态剪枝，实现了基于车辆运动风险控制的渐进优化式局部轨迹规划。最后，采用仿真与实车试验方法验证了基于环境态势场的全局路径和局部轨迹规划算法，试验结果验证了本文提出的环境态势场模型和运动规划方法在越野环境下的安全性、适应性和可行性。

There are many types of obstacles, terrain, moving vehicles, environmental threats in different states and forms in the off-road environment, which affect the vehicle’s motion safety. Planning safe, efficient, feasible path and trajectory in the complex off-road environment is the key technology of unmanned ground vehicles. The traditional vehicle motion planning methods take environmental obstacle avoidance, passenger’s comfort, energy conservation and emission reduction as the objectives, and lack effective evaluation methods for a variety of complex off-road terrain, environmental threats and vehicle risks. It is difficult to balance the safety and traveling efficiency of the path, and could not solve the problem of contradiction between planning time and trajectory optimization during the process of planning. To solve these problems, this study establishes a vehicle motion planning framework based on environmental risk assessment under off-road environment, and presents a unified off-road environmental situation quantitative risk assessment method based on the field theory. A global path planning method of multiple objective balanced probabilistic roadmap and local trajectory planning method of asymptotic optimal rapidly random-exploring tree are proposed based on the environmental situation field model, which plan safe, efficient, feasible path and trajectory in the complex off-road environment.Firstly, a quantitative risk assessment method of off-road environmental situation which considers the effects of multiple environmental elements is proposed. In view of the different types of environmental boundary, static obstacles, terrain, vehicles and environmental threats, which are difficult to be quantitatively evaluated. A unified environmental situation field model is established according to their respective characteristics and status patterns, and the effects of multiple elements in the off-road environment are quantitatively evaluated with the field intensity, working range and direction.Secondly, a global path planning method of multiple objective balanced probabilistic roadmap based on environmental situation field is proposed. The spatial topology map is formed by probabilistic sampling method. The multi-dimensional traffic evaluation matrices are established by evaluating the feasibility, traveling distance and environmental situation of node connection in the topology map. The evaluation matrices are used to analyze and calculate the traveling cost of the vehicle’s planning path. Thus, a global optimal path planning is fulfilled considering the vehicle's off-road riding performance, protection performance and task requirements.Then, a local trajectory planning method of asymptotic optimal rapidly random-exploring tree based on environmental situation field is proposed. The spatial random search tree is generated by the incremental sampling method, the exploring tree trajectory is extended based on the vehicle kinematics constraints, the vehicle driving risk on the motion trajectory is evaluated by the environmental situation field model, and the motion efficiency is evaluated by the vehicle’s motion trajectory length and yaw angle. An asymptotically optimal local trajectory planning is fulfilled by a dynamic trajectory reconstruction.Finally, Simulation experiments and vehicle experiment are carried out to verify the global path and local trajectory planning algorithm based on environmental situation. The experimental results indicates that proposed algorithms are capable of planning a safe, efficient and feasible path in off-road environment.