CoopSim

From ISLAB/CAISR
Revision as of 12:46, 7 March 2014 by 10.0.2.2 (Talk)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Title CoopSim
Summary Simulation of cooperative systems behavior in the presence of faults
Keywords
TimeFrame
References
Prerequisites
Author
Supervisor Stefan Byttner, Cristofer Englund
Level Master
Status Open


Cooperative vehicles with varying degree of automation will be introduced to the traffic system within the coming years. To be able to safely introduce these systems, rigorous safety measures need to be performed. Tests regarding conformance, reliability, robustness to noise, repeatability will be important to perform before vehicles with these systems can be allowed to occupy our roads. Numerous errors originating from sensors, communication, in-vehicle networks, other vehicles, controllers, actuators, and the environment can be present. To cope with such errors, robust systems need to be built. Furthermore, as long as humans drive vehicles within the traffic system, also human errors can be present and need to be handled. This project aims at using VSimRTI , a free traffic and communication simulator software available through FraunHofer Focus in Berlin, to simulate a semi-automated traffic system. The simulator results can be used to evaluate the implemented cooperative functions (application that can help vehicles to coordinate in order to improve safety, traffic flow and/or efficiency). The cooperative function could for example be a cooperative adaptive cruise control (CACC). To make a reliable simulation the traffic flow can be generated by using data from traffic flow counters from the simulated traffic system. By applying an already defined cooperative error model to the simulated traffic system we want to show how the error-model can be used to identify human and machine/environmental errors within the traffic system. The error-model will enable the possibility to create a credibility score for the human drivers in the cooperative system, indicating how trustworthy a driver is. There are a number of underlying questions that we may address in this project namely: Are there errors (human behaviour) that should lower the credibility score more than other errors? Are there errors that force the cooperative functionality to terminate?

Project headlines

  1. Create a simulated traffic environment using VSimRTI
  2. Create vehicle models with cooperative capabilities
  3. Design cooperative functions where errors can be introduced
  4. Compare simulation results of traffic system with/different degree of penetration/without the cooperative function
  5. Introduce errors (human/environment-based errors) in cooperative functions and analyse the effects using the already defined error-model
  6. Develop and estimate a credibility score for the drivers generating unanticipated actions that are classified as errors by the error-model
  7. Report and paper writing

Skills in Java is required.