About Kun’s research
The accident statistics results provided by the Transport Accident Commission (TAC) show that Distracted Driving is one of the major causes of accidents. “I think it is my duty to reduce this risk as a researcher, said Zou.
He is trying to use the power of deep learning and AI technology to develop a robust driver monitoring system for road safety. This system aims to detect the risk behaviours of drivers including playing phone, distraction, and drowsiness. It can also alert the drivers if necessary.
His team has collected the video data of several volunteers in the driving simulator. These data have been used to develop a Convolution Neural Network (CNN) model to detect different behaviours of the drivers including playing phone, distraction, and drowsiness.
This detection system can prevent distracted/drowsy driving and reduce the risk of car accidents. In the future, this system shall work in the autonomous vehicle not only for distraction monitoring but also for monitoring the health conditions of the drivers, so that we can save more drivers if they have acute health problems (such as heart attack, cerebral hemorrhage, or epilepsy).
Reflections on his PhD
The latest knowledge and capability are always needed for developing the latest technology. Kun has learnt how to utilize those new technology and products. He says that searching for the latest papers and updating your think tank frequently are always helpful to get the edge of research.
His supervisors helped him a lot during his research, and he developed my communication skills during his research. For instance, he developed an EEG measurement system and multi-angle cameras based on open-sourced system for his research.
He is very interested in the technology of the digital twin and autonomous driving, and hopes to expand his knowledge to those areas to develop a smarter city in the future.
A word from the PhD supervisor
More than 50% of vehicle accidents are due to a loss of attention caused by distractions such as using a mobile phone, excessive alcohol or drug consumption, and drowsiness. However, the existing driver state monitoring methods are reactive and have limited accuracy/reliability when detecting distraction. Proactive monitoring and warning of driver distractions can significantly reduce the road toll and promote safety in urban and country roads.
It is known that AI can be used to detect distractions, but the reliability and accuracy of detection are still challenging. This project aims to increase the accuracy of the deep learning model and reduce the required computational resources.
This technology can be applied in the automotive industry and can also be used in the aerospace and rail vehicle industry.
Mohammad Fard, RMIT
Contact Kun Zou
If you’d like to contact Kun Zou about his research, please click the button below.