Radio transponders could save lives
Within the COMPOSE subproject, one of the task of Fraunhofer IMS was to investigate together with BMW Research and Technology the feasibility of radio-based detection and localisation of vulnerable road users, such as pedestrians and cyclists, in urban environments. With regard to improved protection, the relevant scenarios were situations in which no line of sight exists between the vehicle and the VRU. In contrast with other sensor systems, e.g. radar or video, a radio-based localisation system is seen as another promising candidate for the detection of hidden moving object.
The system uses active transponders on the VRUs (for example in-built in a mobile phone) and a smart sensing antenna device in the vehicles. The smart antenna device estimates the direction of arrival (DOA) through appropriate algorithms and calculates the subsequent location. This system uses a unique signature code for each VRU device. The unique signature is encoded within the transmitter signal, thus the system does not have the traditional problem of separating different users.
Building on previous work, this study proposes a novel system concept, defines requirements and constrains, and focuses onto the possibility to estimate the DOA of the incident waves. Several problems were addressed :
- Does an appropriate wavelength promise to "look through" obstacles?
- In urban environments, the wave propagates over multiple paths; therefore the smart antenna device “sees” not only signals over the direct path, but additional incident waves from different DOAs
- The formulation of the classical DOA estimation problem requires that the number of impinging signals is less than the number of sensors in the array
- “Clustering” effects
- DOA algorithms, using spatial distributed source (SDS) models to handle the above problems
- Performance comparisons
- Estimation of the spatial POSITION of the VRU transmitter using sequential triangulation, since no direct measurements of distances are available
- Position accuracy estimations
Furthermore the study proposes a realisation concept for the realisation of RF based detection and location.
The study comes to the conclusion, that the feasibility is dependent on the requirements of the driver assistant system (DAS). If the DAS requires only a rough localisation, a realisation could be possible. But if the DAS layer requires also highly accurate information about trajectories of every detectable VRU, then it may be very difficult to reach the requirements. In particular, the accuracy is degraded due to the impossibility to measure the velocity directly and therefore has to be derived from position estimations. The accuracy of the position estimations in turn, are degraded by the fact that the distance is also not accessible directly and has to be calculated by sequential triangulation, using the estimated DOAs.
For further information regarding radio-based detection and the COMPOSE subproject, please contact the subproject leader Marc Walessa or check the subproject website.