Affiliations: [a]
Department of Electronics and Communications Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India
| [b]
Center for Computational Engineering and Networking (CEN), Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore, India
Correspondence:
[*]
Corresponding author. G.A. Shanmugha Sundaram, SIERS Research Laboratory, ASE Coimbatore, Amrita University,
India. Tel.: +91 422 268 5000; Fax: +91 422 268 6274; E-mail: ga_ssundaram@cb.amrita.edu.
Abstract: Real-time occupant posture tracking information has significant value in ADAS equipped vehicles to enable better safety and mitigate risks of injuries to occupants within vehicular cabin during sudden deceleration due to abrupt braking maneuvers or crash scenarios. This information will be helpful for timely activation of airbags and various other conventional safety restraints that provide safety and mitigate injuries. Here, a proximity sensing system is proposed that uses capacitive electrodes to acquire the occupant posture and motion data. These electrodes are deployed as an array placed along three orthogonal sensing axes such as in the seats, along the roof and dashboard, and along the door panel. Motion detection cameras and other sensors like ultrasound or infrared sensors have line of sight issue, while the accumulation of dirt would become a problem for sensing the data accurately. A prototype hardware has been implemented and the proximity capacitance data was acquired for discrete distances from 0.1 to 0.8 m, in the three electrode orientations. Applying optimization and curve fitting techniques on this data, derived data sets were then obtained, that mimic a typical crash-test dummy behavior during impact. The resultant algorithm can offer a precise localization estimate of the occupant with respect to an electrode layout along the roof, seat and door orientations, and hence classify the occupant posture inside the vehicular cabin.
