Background:
Robotic path planning is an important facet of robotics. Its purpose is to
make robots move independently in their work environment from a source to a destination whilst
satisfying certain constraints. Constraint conditions are as follows: avoiding collision with obstacles,
staying as far as possible from the obstacles, traversing the shortest path, taking minimum time,
consuming minimum energy and so on. Hence, the robotic path planning problem is a conditional
constraint optimization problem.
Methods:
To overcome this problem, the Flower Pollination Algorithm, which is a metaheuristic
approach is employed. The effectiveness of Flower Pollination Algorithm is showcased by using
diverse maps. These maps are composed of several fixed obstacles in different positions, a source
and a target position. Initially, the pollinators carrying pollen (candidate solutions) are at the source
location. Subsequently, the pollinators must pave a way towards the target location while simultaneously
averting any obstacles that are encountered enroute. The pollinators should also do so with the
minimum cost possible in terms of distance. The performance of the algorithm in terms of CPU time
is evaluated. Flower Pollination Algorithm was also compared to the Particle Swarm Optimization
algorithm and Ant Colony Optimization algorithm.
Results:
It was observed that Flower Pollination Algorithm is faster than Particle Swarm Optimization
and Ant Colony Optimization in terms of CPU time for the same number of iterations to find an
optimized solution for robotic path planning.
Conclusion:
The Flower Pollination Algorithm can be effectively applied for solving robotic path
planning problem with static obstacles.