Detect: Probabilistic information fusion
To locate targets and trace them, the patented "Bayesian Occupancy filter" (BOF) merges several data sources. Thus the field of action is extended while better location and reduction of target losses are ensured.
This software technology has been used successfully with sound sensors, images and radars. It increases the efficiency of systems without the need to modify hardware architectures.
Examples of achievements:
- Fusion of Video and Lidar sensors to produce qualitative information (pedestrians, vehicles, animals) and quantitative information (distance, accurate speed).
- Networking of sensors to extend the monitored area, improve accuracy and eliminate errors in the overlap areas.
Predict: Behaviour prediction
The learning capabilities of our probabilistic technologies will allow you, just by supplying examples, to program the typical behaviours that have to be detected and add knowledge if applicable (formal models or expert rules). In operational mode, the software computes in real time the probability of occurrence of each behaviour as well as its probable change.
Examples of achievements:
- Smart autonomous system, programmed by behaviour learning.
- Classification, identification and prediction of threatening behaviours for anti-aircraft defence.
- Prediction and computation of risks of collision for automatic operation of vehicles.
