Signature-based multi-tracking system
on reconfigurable multi-core architecture
Increasing levels of complexity and computational demands in automotive applications forced a move to powerful yet cost-effective embedded system.
Our recent research activities concentrate on investigating the design of driver assistant system in complex driving scenarios, such as detecting pedestrians or under changing weather and lighting conditions.
However, existing driver assistance systems do not support these complex applications because of their limited functionalities and cost constraints for large-scale automotive use.
The work to be done in this post-doctorate position aims to contribute in the design of an early warning and collision avoidance system. During the last 3 years, our group succeeded in developing two systems for DAS : the recognition system that is able to categorize the obstacles using a complex and expensive obstacle radar signature (ORS) in one hand and the multiple target tracking system (MTT) that is able to track the obstacle positions but with out taking into account their signatures.
The proposal here is to come up with a full system where the ORS and the MTT work in a coordinated manner. This cooperation between the 2 parts is very effective as it may reduce system cost and enhance accuracy of the DAS.
The job include :
- designing an efficient cooperation methodology and approach for cooperation and data exchange between ORS and MTT,
- designing an efficient multi-core architecture on FGPGA.
Contacts
smail niar @univ-valenciennes.fr
Employment
Gross monthly wages : 2500 €
Duration :12 months
Employer : CNRS France
*********************
Pr. Smail NIAR
University of Valenciennes (France)
email:
tel: +33 327511948
Mobile: +33 688496057
*************************
--
Best Regards,
PhD Student
Circuits and Systems Group
Electrical and Electronic Engineering
Imperial College
South Kensington Campus
London SW7 2AZ
__._,_.___
MARKETPLACE
.
__,_._,___
No comments:
Post a Comment