Mobile robot optimum trajectory Development using a hybrid reactive navigation model

Abstract

Path planning for mobile robot navigation in workspaces with varying obstacles complexity levels was addressed in this research. The domain problem is that for a specific class of obstacles referred to as the concave shaped and lengthy obstacles, the likelihood of local minima trap occurring is often significantly high. For instance, a labyrinth premised on concave shaped obstacles often misleads a navigating robot into the concave hollow region in a bid for the robot to reach its desired target point. Apart from the use of reactive algorithms, for an autonomous navigation process which is often premised on continuous path trajectory development, the literature clearly alleges that most non-reactive algorithms get trapped in the concave hollow and along the edges of lengthy obstacles. The purpose of this research is to adapt a reactive mobile robot (MR) navigation algorithm premised on the Hybrid Virtual Force Field (HVFF) concept for the exploration of robot navigation in both developed and literature based obstacle constrained workspaces. The obstacles considered in this research work are mostly premised on concave shaped and lengthy obstacles cul-de-sac. The HVFF approach evolved from the Virtual Force Field (VFF) approach which is premised on the Potential Field Method (PFM). This method of path planning operates by utilizing the resultant of forces emanating from the combination of repulsive and attractive forces acting on a navigating robot. The algorithmic validation was carried out via the conduct of simulation trials using the Python software. The simulations conducted span across newly developed workspaces and literature based workspaces for a comparative study. Furthermore, the behaviour of the robot navigation with and without the HVFF algorithm per workspace was presented. Of a particular interest was the navigation time with and without the HVFF algorithm per workspace. The results obtained in all the simulations showed a much efficient navigation completion time with the use of the HVFF algorithm. Efficiency in arriving at the target point implies that the robot was able to come out of the local minima trap each time it entered the hollow region of a concave shaped obstacle or around the edges of a lengthy stretched out obstacle. The time difference recorded between deploying the HVFF approach and not deploying the HVFF algorithm across the different simulations conducted spanned between 14.27 to 287.44 seconds which corresponds to a percentage gain time of 31.87% and 89.70% including a simulation with an unending target point (TP) arrival time for the without HVFF algorithm. As the concave trap increased in its depth, the tendency of the robot to escape from the trap becomes much more difficult. The outputs of this research justify the effectiveness and efficiency of the HVFF algorithm.