Methodology of Using Terrain Passability Maps for Planning the Movement of Troops and Navigation of Unmanned Ground Vehicles

Abstract

The determination of the route of movement is a key factor which enables navigation. In this article, the authors present the methodology of using different resolution terrain passability maps to generate graphs, which allow for the determination of the optimal route between two points. The routes are generated with the use of two commonly used pathfinding algorithms: Dijkstra’s and A-star. The proposed methodology allows for the determination of routes in various variants—a more secure route that avoids all terrain obstacles with a wide curve, or a shorter route, which is, however, more difficult to pass. In order to achieve that, two functions that modify the value of the index of passability (IOP), which is assigned to the primary fields that the passability map consists of, have been used. These functions have a β parameter that augments or reduces the impact of the applied function on IOP values. The paper also shows the possibilities of implementation of the methodology for the movement of single vehicles or unmanned ground vehicles (UGVs) by using detailed maps as well as for determining routes for large military operational units moving in a 1 km wide corridor. The obtained results show that the change in β value causes the change of a course of the route as expected and that Dijkstra’s algorithm is more stable and slightly faster than A-star. The area of application of the presented methodology is very wide because, except for planning the movement of unmanned ground vehicles or military units of different sizes, it can be used in crisis management, where the possibility of reaching the area outside the road network can be of key importance for the success of the salvage operation.

A Hybrid Multi-Target Path Planning Algorithm for Unmanned Cruise Ship in an Unknown Obstacle Environment

“…Since the 1970s, many studies on the path planning problem have been conducted. The path planning methods can be roughly divided into several groups: the grid search methods, such as A* algorithm [ 2 ], Depth-First Search (DFS) [ 3 ], Breadth-first Search (BFS) [ 4 ], and Dijkstra algorithm [ 5 ]; the sampling-based methods, such as Probabilistic Roadmap (PRM) [ 6 ] and Rapidly Exploring Random Tree (RRT) [ 7 ]; heuristic or swarm intelligence algorithms, such as Genetic Algorithm (GA) [ 8 ], Ant Colony Optimization (ACO) [ 9 ], Particle Swarm Optimization (PSO) [ 10 ], and neural network-based algorithms [ 11 ]; the potential field methods, such as Artificial Potential Field (APF) [ 12 ], optimal-control based method [ 13 ], and geometry-based method [ 14 ]. The listed algorithms have certain advantages and disadvantages.…”

Fast Route Planner Considering Terrain Information

“…There are many papers on route planning considering terrain information [8][9][10][11][12]. The authors of [8] presented a control architecture for fast quadruped locomotion over rough terrain.…”

“…[8] applied a Dynamic Programming (DP) algorithm to plan a minimum-cost path across the terrain. The authors of [9] addressed using different resolution terrain passability maps to construct graphs, which allow for the determination of the optimal route between two points. The routes were generated using two path planners: Dijkstra's and A * .…”

“…However, DP [8], A * [9,11,12], or D * [10] require that the entire region is already completely covered by multiple grid cells. As one increases the entire workspace size, the number of grid cells covering the workspace also increases.…”

“…As one increases the size of the entire region in ACO [6], DP [8], A * [9,11,12], D * [10], or Theta * [4,5], the number of grid cells covering the entire region also increases. Our paper does not use grid cells to cover the entire region, and the shortest route is built using the node network deployed by two virtual vehicles.…”