From Sensor-Space to Eigenspace – A Novel Real-Time Obstacle Avoidance Method for Mobile Robots

Abstract

This paper presents a novel real-time obstacle avoidance and navigation technique called as “Free-configuration Eigenspace” (FCE). The FCE enables an autonomous robot to detect unknown obstacles and avoid collisions while simultaneously steering the robot towards the target. The methodology utilizes a two-dimensional (2D) Cartesian Eigenspace as a world model. 2D way points (in world model) are extracted by computing (through FCE model) stacked Eigenvectors of laser data at discrete time scans which manifest desired robotic trajectory. The world model is updated continuously upon obtaining discrete time laser scans sampled by on-board Light Detection and Ranging (LiDAR) sensors. The FCE technique has been implemented on Robotic Operating System and real-time robotics simulator Gazebo${}^{\circledR }$. Encouraging results are obtained and shown in this paper. The FCE has also been tested in real-time with laser scans obtained from VL6180X LiDAR sensor mounted on a custom autonomous robot. The results obtained are further compared with the state-of-the-art obstacle avoidance and path planning method called as vector field histogram (VFH). The FCE results (shown in the paper) are comparable, encouraging, and outperform VFH in path length and safe distance maintenance.

Keywords:

• Autonomous robots
• Eigenspace
• Eigenvector
• Path-planning
• Robotic navigation

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Shyba Zaheer

Shyba Zaheer received the MTech degree in computer science and data processing from Indian Institute of Technology (IIT), Kharagpur, India in 2003, Bachelor of Technology in electrical and electronics engineering, Kerala University, India in 1991. Currently working as faculty in department of Electrical and Electronics Engineering, TKM College of Engineering, Kollam, Kerala, India, pursuing PhD in robotics and artificial intelligence. Her main research interests are robotics, artificial intelligence, and signal processing. Corresponding author. Email: [email protected]

Tauseef Gulrez

Tauseef Gulrez received the master's degree in computer systems engineering from the University of Technology, Sydney, Australia, in 2005, studied neuro-engineering as a predoctoral research fellow from the Rehabilitation Institute of Chicago, Northwestern University, Chicago, IL, USA, in 2006, and received the PhD degree in robotics and computer science from Macquarie University, Sydney, in 2008. He was a research fellow at Mechatronics and Haptic Interface Lab, School of Engineering, Rice University, Houston, TX, USA, and Learning and Affect Technologies, The University of Sydney, Sydney. He held positions of senior researcher at Smart Medical Devices Laboratory, Qatar University and at the Center for Robotics and Autonomous Systems, University of Salford, Manchester, United Kingdom. Presently, he is a collaborator researcher with Virtual and Simulations of Reality, Lab, Department of Computing, Macquarie University, Sydney, Australia. He is an author of a book and more than 50 research journal and conference articles. His main research interests include robotics, tactile feedback user interfaces, virtual reality systems, and signal processing. Email: [email protected]

Imthias Ahamed Thythodath Paramabath

Imthias Ahamed Thythodath Paramabath received his BTech degree in electrical engineering from Kerala university and MTech degree in instrumentation and control from NIT-Calicut, India. He obtained his PhD from Indian Institute of Science, Bangalore, India in 2002. He has more than 25 years of teaching and research experience which include 3 years in Kind Saud University, Riyadh, Saudi Arabia, and 2 years in Dhofar University, Salalah, Oman. He is currently a professor with the Department of Electrical and Electronics Engineerng, TKM College of Engineering, Kollam. He has published three books and few papers. His current research include reinforcement learning, robotics, neural networks and power system scheduling and control. Email: [email protected]