A novel approach to driver behavior prediction using scene context and physical evidence for intelligent Adaptive Cruise Control (i-ACC)

Conventional driver assistance systems suffer from relatively late reaction timing. Adaptive Cruise Control (ACC) systems, for example, react to vehicles cutting-in from neighboring lanes once these vehicles are at least partly driving on the host vehicle's lane. The improvement by state of the art prediction approaches is limited, because the reliable approaches base on features, which are characteristic for the phenomenological effect of a behavior change. That means they are capable of predicting a behavior once it started. The Honda i-ACC overcomes the above mentioned limitation of delayed reaction by using a novel approach of behavior prediction. This approach combines two prediction methods: A context-based prediction (CBP) and a physical prediction (PP). The CBP evaluates the situational context of a predicted vehicle and does not use any phenomenological feature. It is thus capable of long-term prediction. In contrast, the PP evaluates phenomenological features by accumulating the predicted vehicle's history of recent positions and comparing them to a set of trajectories, which allows for an accurate short-term prediction. In an evaluation on 15000km of driving in Europe we will show that the new prediction approach achieves both, reliable and long-term prediction.

• Record URL:
  http://publications.lib.chalmers.se/records/fulltext/222422/local_222422.pdf
• 
• Corporate Authors:

  Chalmers University of Technology. SAFER Vehicle and Traffic Safety Centre

  ,    
• Authors:
  • Schmuedderich, Jens
  • Rebhan, Sven
  • Weisswange, Thomas
  • Kleinehagenbrock, Marcus
  • Kastner, Robert
  • Nishigaki, Morimichi
  • Kamiya, Hiroyuki
  • Mori, Naoki
  • Kusuhara, Shunsuke
  • Ishida, Shinnosuke
• Conference:
  • FAST-zero'15: 3rd International Symposium on Future Active Safety Technology Toward zero traffic accidents, 2015
  • Location: Gothenburg , Sweden
  • Date: 2015-9-9 to 2015-9-11
• Publication Date: 2015

Language

• English

Media Info

• Pagination: pp 85-92
• Monograph Title: FAST-zero'15: 3rd international symposium on future active safety technology toward zero traffic accidents: September 9-11, 2015 Gothenburg, Sweden: proceedings

Subject/Index Terms

• TRT Terms: Autonomous intelligent cruise control; Detection and identification; Driver support systems; Traffic; Vehicles
• ITRD Terms: 1451: Adaptive cruise control; 8743: ADAS; 8743: Advanced driver assistance system; 9115: Detection; 8743: Driving aid (electronic); 8743: Electronic driving aid; 8743: Electronic trafic aid; 755: Traffic; 1255: Vehicle
• Subject Areas: Operations and Traffic Management; Vehicles and Equipment; I91: Vehicle Design and Safety;

Filing Info

• Accession Number: 01602265
• Record Type: Publication
• Source Agency: Swedish National Road and Transport Research Institute (VTI)
• Files: ITRD, VTI
• Created Date: Jun 20 2016 1:26PM