Follow Us on Google
A Nigerian robotics researcher, Otobong Jerome, has developed a navigation system designed to help wheeled robots move safely across rough and uneven terrain, including ploughed fields, slopes and broken ground.
Jerome, who works at the Centre for Autonomous Technologies, Innopolis University, developed the method with Alexandr Klimchik, Alexander Maloletov and Geesara Kulathunga of Innopolis University and the University of Lincoln, United Kingdom.
The research, published in the IEEE Robotics and Automation Letters, focuses on one of the major challenges facing the use of autonomous machines in agriculture and other difficult environments.
Unlike warehouses, factories and city roads, where many existing navigation systems operate, farmland is often characterised by ruts, steep slopes and unstable surfaces, making it difficult for driverless vehicles to determine safe routes.
Jerome and his collaborators developed a navigation method based on a genetic algorithm, which generates several possible driving routes, evaluates them and repeatedly combines and modifies the best options until it identifies a suitable path.
The system does not simply search for the shortest route. It also considers the physical limitations of the vehicle and the difficulty of the terrain, avoiding steep climbs and sudden changes in slope that could cause a vehicle to lose stability or overturn.
According to the research, the method produced gentler routes across different challenging terrains while keeping the additional distance travelled relatively low. It also performed more consistently than competing navigation methods tested by the researchers.
The team further tested the system on a small four-wheeled robot at Innopolis University, where the machine used sensors to build a three-dimensional picture of its surroundings before planning and following a route across outdoor terrain.
The researchers said the technology could have applications beyond agriculture, including autonomous vehicles used in mines and quarries, disaster-response robots and planetary exploration rovers. The team has also made the software openly available to enable other researchers to adapt it to different machines and missions.
Jerome and his collaborators are now exploring further improvements, including methods that could allow autonomous vehicles to learn their own behaviour while operating. The research highlights the growing role of robotics and artificial intelligence in developing machines capable of safely operating in environments where conventional navigation systems struggle.
