DC5 Project: Enhancing Perception Accuracy through Collaborative Sensor Fusion Framework in Autonomous Systems

PhD Supervisor: prof. Jerry C.-W. Lin (SUT); Auxiliary supervisors: Jia-Chun Lin (NTNU); Vicente García Díaz UNIOVI), Volker Stolz (HVL); R&D cooperation: Marek Drewniak (AIUT)

 

Objectives: By improving the accuracy of perception, the designed system will better recognise and identify objects, obstacles, road conditions and other relevant information. This in turn will enable better decision making and navigation. By providing high-quality, real-time data about the environment, the designed system will also make better decisions, adapt to changing conditions and operate effectively in complex and dynamic scenarios. The framework is also designed to be adaptable, thus enabling autonomous systems to adapt to different environmental and operational conditions. This flexibility is crucial for systems that need to operate in different environments, from urban streets to off-road terrain. The framework should seek to optimise the use of sensors and computational resources in order to achieve a high degree of accuracy while minimising energy consumption and computational load. The framework is designed to accommodate different sensors and configurations. This flexibility will enable developers to select the most suitable sensors for their specific application.

 

Expected Results: The framework should lead to better object detection capabilities. This means that autonomous systems will be able to identify and track objects with greater accuracy. This is crucial for safe and efficient operational decisions. With collaborative sensor fusion, the system will also have built-in redundancy. If one sensor fails or provides faulty data, other sensors will step in to ensure continuous operation and safety. The resource optimisation of the framework should lead to a more efficient use of computing power and energy. This is particularly important for autonomous systems that run on battery power as it can increase their range and reduce operating costs. The ability to integrate a wide range of sensor types and configurations will ensure that the framework is able to adapt to different applications and requirements. This scalability will make it possible to customise autonomous systems for specific use cases while also improving perception accuracy. Finally, the implementation of this framework is expected to drive technological advances in the areas of sensor technologies, signal processing and artificial intelligence. This will have a far-reaching impact on the field of robotics and autonomous systems beyond its immediate application.

 

Applied research: The research results on design approaches, development and improving the accuracy of perception will be verified by AIUT in its industrial applications. Newly developed methods for sensor fusion and signal processing can potentially extend portfolio of technology used by AIUT.

 

Planned secondments: NTNU (4 months); UNIOVI(4 months); HVL (4 months)                    

 

Enrolment in Doctoral degree: SUT

 

Recruitment of Doctoral Candidate: https://euraxess.ec.europa.eu/jobs/28797