BOF

Bayesian Occupancy Filter for robust sensing and multi-target tracking

In dynamic environments, the ability to detect moving objects, to estimate their positions, and to predict their motion is central for a large category of applications:

 Driving assistance and safety,

 Autonomous robots,

 Sensor-based surveillance,

 Smart interactive toys.

The Bayesian Occupancy Filter (BOF) technology provides a strong theoretical framework for robust sensing in dynamic environments using imperfect data coming from various sensors. By collecting sensors input information, it uses the priors on the behaviors of surrounding obstacles to maintain a stabilized map of their locations and velocities. The BOF technology considerably reduces the impact of noise and false alarms in scene understanding while allowing robust sensor fusion by performing a spatial and temporal filtering of the input readings.

Probayes develops the BOF++ library as an efficient and robust implementation of the BOF framework. This multiplatform object-oriented C++ library has been designed to ensure:

 Very fast computation allowing real-time use,

 Extensibility by using "pluggable" occupancy and velocity sensors.

Probayes capitalizes a long experience and an unsurpassed know-how in using the BOF technology in real-world environments and using various sensor modalities including:

 Computer vision-based detectors,

 Laser range finders,

 Stereovision-based 3D sensors.