106 Int. J. Bio-Inspired Computation, Vol. 9, No. 2, 2017
Copyright © 2017 Inderscience Enterprises Ltd.
Mobile robot path planning with modified ant
colony optimisation
Utkarsh Rajput* and Madhu Kumari
Computer Science and Engineering Department,
National Institute of Technology,
Hamirpur-177005, Himachal Pradesh, India
Email: utkarshrjpt@gmail.com
Email: madhu.jaglan@gmail.com
Email: madhu@nith.ac.in
*Corresponding author
Abstract: Automated navigation is a pivotal task of robotics research and the key challenge lies
in robot motion on unknown dynamic terrain. The large number of solutions to robotic path
planning, especially in unknown and dynamic environments, mainly rely on the heuristic
methods. The most important factor for this choice is the fast convergence towards solution
without supervision. In the proposed scheme we have used a modified version of ant colony
optimisation. We incorporated the directional movement history of robot on a grid into a vector
as a probability multiplication factor which helps to achieve a faster convergence and avoid
unnecessary movements, e.g., looping. In this work we have devised a novel pheromone updation
scheme. Along with this we have applied path smoothing to lessen the number of turns on the
candidate optimal path. Effectiveness is shown through several extensive experiments and results
clearly indicate the aptness of the proposed scheme.
Keywords: mobile robot; ant colony optimisation; ACO; path planning; path smoothing.
Reference to this paper should be made as follows: Rajput, U. and Kumari, M. (2017) ‘Mobile
robot path planning with modified ant colony optimisation’, Int. J. Bio-Inspired Computation,
Vol. 9, No. 2, pp.106–113.
Biographical notes: Utkarsh Rajput received his Bachelor of Technology in 2012 from Gautam
Buddh Technical University, Lucknow, Uttar Pradesh, India. He received his Master’s from
National Institute of Technology, Hamirpur, Himachal Pradesh, India. Machine learning and
robotics are his major interest areas.
Madhu Kumari is currently working as an Assistant Professor in the Department of Computer
Science and Engineering at National Institute of Technology, Hamirpur, Himachal Pradesh,
India. She received her Master’s and doctoral degrees from Jawaharlal Nehru University, New
Delhi. Her previous work mostly focused on exploration of different variants of reinforcement
learning in simulated Robocop soccer domain. She has worked on different aspects of
computational advertising based on sponsored search auctions. Her current research is more
aligned towards the deployment of machine learning techniques on application milieu like web,
text and social analytics.
1 Introduction
Mobile robots have their applications in diverse fields
ranging from dangerous environments, like mining and
nuclear industry to daily life applications, e.g., autopilot
robots for cars and household work. These robots can be
successfully used in hazardous environments where human
life can be endangered. For the development of mobile
robots, effective and efficient path planning methods are
necessary for the locomotion. Robot path planning is a
pivotal research problem and researchers are continuously
trying to develop methods that can be used an integrable
solution for this problem in a manner that can find the
optimal path in real time with least number of constraints.
But there has been less success achieved in this area.
Initially the main focus was on the path planning in static
environments where the obstacles are static and robot is the
only mobile entity. Major techniques which have been
extensively used in static environments based path planning
includes cell decomposition method (Choset, 2007),
visibility graph method (Mitchell, 1986), Voronoi diagrams
(Kalra et al., 2009), etc. In cell decomposition method
(Choset, 2007) environment map is decomposed into
different sized cells labelled as obstacle and free cell. A
possible navigation route for robot is through connected free
cells only. The incremental algorithm in Kalra et al. (2009)
uses techniques for constructing and updating generalised
Voronoi diagrams and thus finding the route through
unknown static environment. Genetic algorithm with