Path Planning Of Unmanned Ground Vehicle In Static And Dynamic Constraints Environment

In the recent era,implementation of Unmanned Ground Vehicle(UGV)increases in different areas including disaster management,Industry,military,and function ranging from the applications to assist human beings in daily life,improve their safety and comfort.In this scenario,UGV needs to further improve its capabilities of intelligent decision making to accomplish desired tasks smoothly in static and dynamic constraints environment.In the content of this,path planning is considered as a challenging task in motion planning of UGV.Based on environment complexity,path planning has been categorized into two approaches known as Global Path Planning(GPP)and Local Path Planning(LPP).GPP approach generates a suitable path for vehicles from the start position to destination in a static constraint environment,meanwhile many unexpected constraints may appear on the way towards the destination in front of the vehicle.To avoid those obstacles and keep moving on reference global trajectory,the LPP approach is applied.This thesis provides a collective approach to path planning by generating a smooth global route with local planning ability to avoid unexpected obstacles in maps of different complexity with improving safety,path consistency,and trajectory planning for UGVs.This approach has been derived to overcome the deficiencies of path planning in a grid-based environment.1.Ant Colony Optimization(ACO)is an evolutionary approach.Due to features such as strong robustness,good feedback information,and better-distributed computing,it has become a famous approach and it can be easily combined with different techniques to deliver the desired optimal solutions.In this thesis,an efficient approach has been derived to achieve optimal trajectory in complex environments.A novel concept of using A* Multi-Directional algorithm with heuristic features has been introduced to enhance the computational efficiency of the Ant Colony Algorithm(ACO).In the first part,A* Multi-directional algorithm starts to search in the map and stores the best close nodes between start and destination with optimal heuristic value,and that area of nodes has been chosen for path search by ACO.It assists ACO to avoid blind search and direct it towards the final goal.A novel evaluation model is introduced with an optimal reward policy to reduce the computational complexity of filtering the global waypoints to remove the bending effect.Path obtained from ACO in a grid-based environment is considered a rough path due to the number of bending and poor visibility to localize the UGV.This problem is concerned with most GPP approaches that applied in a grid-based environment.Grid-based representation is a more dominant environment used mostly by global path planners due to its efficient computing ability in computers compared with other environments.Meanwhile,the route generated through a grid-based environment consists of sequences of points called path candidates,which make straight suboptimal line segments containing sharp bending and the result is a rough path.With arc-length parameterization,a curvilinear smooth route has been generated among filtered waypoints.It provides consistency and smoothness in the global reference path to minimize the amount of sharp turning and braking on specific applications.2.The optimal global route has been taken as the centerline of the road and considered as reference route for UGV to navigate in a dynamic constraints environment and localize itself to maintain its offset position near the reference route.3.To achieve comfort drive and safety of robot,lateral and longitudinal control has been utilized to form a set of optimal trajectories along the reference route,as well as,minimizing total cost.Additionally,for collision detection,an optimal local trajectory has been achieved for each forward movement of a robot.The results have been verified through simulations in MATLAB and compared with previous global path planning algorithms to differentiate efficiency and quality of path in static and dynamic constraint environments.