TY - JOUR
T1 - A Dynamic Colour Perception System for Autonomous Robot Navigation on Unmarked Roads
AU - Narayan, Aparajit
AU - Tuci, Elio
AU - Labrosse, Frédéric
AU - Alkilabi, Muhanad Hayder Mohammed
N1 - Funding Information:
Supplementary Data S1 Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/ Supplementary Data S1 Aparajit Narayan obtained his B.Eng. (Hons) in electronic engineering from the University of Sheffield in 2013. He is currently completing his Ph.D. in computer science from Aberystwyth University (UK), working on a project funded by Fujtisu-HPC Wales. His primary research interests are in evolutionary robotics and deep-learning, especially with regards to the application of control strategies emerging from these fields to various real-world problem domains. Dr. Elio Tuci is a Senior Lecturer in the Department of Computer Science, Middlesex University (London, UK). He received a Ph.D. in Computer Science and Artificial Intelligence from the University of Sussex, Brighton, UK in 2003. His research interests fall into the interdisciplinary domain of bio-inspired robotics and computational intelligence, drawing inspiration from nature to design control mechanisms to allow artificial agents to operate in a complex environment and to learn from their experience in an autonomous way. Dr Frédéric Labrosse obtained his degree from Université Paris XI, France, in 1991 and his Ph.D. from École Polytechnique de Montréal, Canada, in 1999. He currently is a Senior Lecturer in Computer Science, Aberystwyth University, UK and his research concerns Robotics and Computer Vision, and in particular the interface between the two fields. As such he has worked on visual sensors and visual navigation methods. He is also interested in the practical applications of robotics, collaborating with many customers who need sensors in remote, dangerous, places. Muhanad Alkilabi obtained his electrical engineering bachelors degree from Al-Mustansiriyah University, Baghdad in 2003 and a Masters degree from Pune University, India in 2007. He is a member of faculty in the Computer Science department at the University of Kerbala and is currently pursuing his Ph.D. in swarm robotics from Aberystwyth University, UK. His research interest is in bio-inspired robotics and computational intelligence.
Funding Information:
This work is partly sponsored by Fujitsu and HPC-Wales.
Publisher Copyright:
© 2017
PY - 2018/1/31
Y1 - 2018/1/31
N2 - Navigation on unmarked and possible poorly delineated roads where the boundaries between the road and the non-road surfaces are not clearly indicated is a particularly challenging task for autonomous vehicles. The results of this study show that fairly robust navigation strategies can be generated by a robot equipped with a dynamic active-vision based control system represented by an artificial neural network synthesized using evolutionary computation techniques. In the experiments described in this paper, a simulated Pioneer robot is required to visually navigate multiple poorly delineated roads that differ in terms of variations in luminance and/or chrominance between the road and the adjacent non-road areas. Low resolution camera images are processed by a mechanism that continuously adjusts the contribution of each component of a three dimensional colour model (e.g., R, G and B) to the generation of the robot perceptual experience. We show that the best controller can successfully drive a simulated Pioneer robot in environments with colour characteristics never encountered during the design phase, and operate with colour models never used during training. We show that the dynamic differential weighting of the colour components is underpinned by a complex pattern of neural activity that allows the robot to successfully adapt its perceptual system to the colour characteristics of different visual scenes. We also show that the controller can be easily ported onto real hardware, by showing the results of a series of tests with a physical Pioneer robot required to navigate various poorly delineated pedestrian roads.
AB - Navigation on unmarked and possible poorly delineated roads where the boundaries between the road and the non-road surfaces are not clearly indicated is a particularly challenging task for autonomous vehicles. The results of this study show that fairly robust navigation strategies can be generated by a robot equipped with a dynamic active-vision based control system represented by an artificial neural network synthesized using evolutionary computation techniques. In the experiments described in this paper, a simulated Pioneer robot is required to visually navigate multiple poorly delineated roads that differ in terms of variations in luminance and/or chrominance between the road and the adjacent non-road areas. Low resolution camera images are processed by a mechanism that continuously adjusts the contribution of each component of a three dimensional colour model (e.g., R, G and B) to the generation of the robot perceptual experience. We show that the best controller can successfully drive a simulated Pioneer robot in environments with colour characteristics never encountered during the design phase, and operate with colour models never used during training. We show that the dynamic differential weighting of the colour components is underpinned by a complex pattern of neural activity that allows the robot to successfully adapt its perceptual system to the colour characteristics of different visual scenes. We also show that the controller can be easily ported onto real hardware, by showing the results of a series of tests with a physical Pioneer robot required to navigate various poorly delineated pedestrian roads.
KW - Active-vision
KW - Evolutionary robotics
KW - Road-following
UR - https://www.sciencedirect.com/science/article/abs/pii/S0925231217317228
UR - http://www.scopus.com/inward/record.url?scp=85033771926&partnerID=8YFLogxK
U2 - 10.1016/j.neucom.2017.11.008
DO - 10.1016/j.neucom.2017.11.008
M3 - Article
SN - 0925-2312
VL - 275
SP - 2251
EP - 2263
JO - Neurocomputing
JF - Neurocomputing
ER -