Self-driving, intelligent cars might soon be able to predict driver distraction thanks to research from leading world universities.
Edge Hill University will monitor driver’s actions to help predict when they might become distracted and unable to take control of the vehicle as required, in situations varying from heavy traffic to bad weather.
Using cameras, they will gather data on the surrounding traffic and in vehicle activities involving driver distraction such as using a mobile phone, eating or getting sleepy as they drive around one of India’s largest cities with a population of over 10 million.
Working alongside experts at India’s top universities, the Indian Institute of Science (IISc), Indian Institute of Technology (IIT), Madras and the University of Bristol, the team will then create algorithms helping intelligent, autonomous cars predict when their drivers might not be able to take control of the vehicle.
Edge Hill’s Dr Ardendu Behera, a Senior Lecturer in Computer Science who is leading the CHARM project (Context-aware Human Activity Recognition and Monitoring for intelligent vehicles), said:
“The aim is for this project to be a clear front-runner in real-time unobtrusive monitoring of complex human activities during driving. Today’s cars are nearly autonomous due to advanced driver assistance systems for safe driving ranging from cruise control to emergency breaking. This is possible thanks to the use of cameras, sensors and other devices which focus on perceiving the surrounding environment.
“However, minimal work has focused on the human driver perspective and human error is overwhelmingly to blame for up to 94 per cent of accidents. Efficient driver monitoring is the bottleneck and being able to track activity inside the car will enable the vehicle to take control during potentially unsafe situations and prepare for any corrective measures such as automatic maneuvering to avoid collision, slowdown and breaking. This is not only crucial for safety and comfort but also key to the transfer of control between driver and vehicle in highly automated driving.
“Traditional human machine interfaces such as touchscreens, buttons and switches will soon be replaced with more natural interfaces such as voice commands, hand gestures and body languages including facial expressions.
“Ultimately CHARM will recognize in-vehicle activity (attentiveness, mobile phones, drinking etc), possible next activities and be able to anticipate context (congestion, cyclist, roadworks etc.), to enable the vehicle to be context-aware.”
“Existing data and algorithms only focus on vehicles driven in Europe and the US. We plan for 15-20 participants to drive around the busy Indian city of Bangalore in various traffic and weather conditions while we monitor their pose, movements and interaction with objects of interest inside the car such as picking up a bottle or touching the dashboard.
“We will then use this data to create algorithms which enable the vehicle to prepare for corrective measures such as automatic breaking and collision avoidance taking input from both external and internal sensors.”
Over the two-year project, CHARM will allow researchers and students to exchange knowledge and visit each country.
CHARM’s progress and research outcomes will be promoted to industry and leading researchers and showcased at workshops, symposiums and conferences worldwide.
The research is jointly funded by the UK-India Education and Research Initiative (UKIERI) and the Department of Science and Technology (DST) of the Government of India.
To find out more the University’s Computer Science department here