A Path Planning Research-Implementation Based On Environmental Reference & Action Element For The Autonomous Vehicle

The planning is the kernel of robot to finish tasks given by users in central decision architecture. The planning and environmental modeling restrict system adaptability to the environment, complex environment handling capacity, as well as expeditiousness and agility of navigating and monitoring. In company with human-interface, it also limits user's acceptation degree to the application systems. The aim of planning is to find the optimal route, action sequence and schedule connection among walking, operating, apperceiving and driving to complete the given task according to the setting estimating criteria. In the experimental system, we get rid of the strict environment recognition demand in central decision architecture, apply quick reaction idea for emergency in action comprehension architecture using planning, environmental reference, sensor data fusing, common sense and experience utilization together. Then, the anti-collision walking can be replied freely in many environmental conditions, for example, completely known, partly known and entirely unknown. In this way, we could not only use central decision and action comprehension together in the engineering systems, but also extend extremely practicality in robotics.In this paper, the author analyzes briefly about puzzles in present robot planning, illustrates the role of action element concept in path planning, discusses deeply about the problems in action element planning and brings out the corresponding solution. After analyzing finely about the design conception, the frame construction, and every module's function partition in the system, we bring forward the technology guideline and implementation strategy according with practical application and feasibility.First of all, the experimental system performs the path planning at physical level based on link knowledge of the environmental regions, gets usable path under the chosen estimating criteria, and enforces path optimization based on simplicity of driving. Then, the trajectory planning at the walking action element level refers to the vehicle characteristicand the task accomplishment satisfaction, may substitute interpolation walking segments for inconsequent or unimplemented ones. According to environmental boundary and region allocation nearby the path, the system establishes velocity at different position in accordance with safety and rapidity, submits the result to the monitoring module. During the system implementation, the author brings forward a complex of many search algorithms, builds up kinematics model of central-position turning vehicle, and puts forward a new path interpolation method. In addition, it also has been researched dynamic timely actions among drive units, discussed failure solving method, improved upon original action element planning method for being applied to all sorts of environmental condition more smartly. Finally, in the paper, it is presented running and testing result of experimental system, pointed out the limitation and insufficiency which still is included, putted forward further ideas and targets under the subject development trend.