BOF doc

BOF Technology

The Bayesian Occupancy Filter (BOF) is a specialized model of the well known Bayesian filtering concept.

In the BOF framework, the environment is represented as a 2-dimensional planar grid. Each cell of the grid contains two probability distributions. A probability distribution on the occupancy of the cell, and the probability distribution on the velocity of the cell occupancy. The probability distributions of the cell occupancy and velocities are estimated by a Bayesian filter given the set of sensor data readings.

Specialized fast clustering/association techniques are then applied on the occupancy/velocity grid in order to track the moving obstacles present in the scene.

The BOF technology is patented by Probayes and INRIA. Probayes owns an exclusive license on the BOF technology.

BOF++ library

The BOF++ library is 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.

This library provides the following set of modules:

• BOF main class allowing to define:
  • Scene geometry and the spatial discretization parameters
  • Velocity discretization parameters
• “Pluggable” sensor models describing:
  • Occupancy sensors
  • Velocity sensors
• Muti-target tracking implementing:
  • Fast clustering algorithms
  • Fast association algorithms
  • Kalman filter updating
• Visualization classes (OpenCV, OpenGL)

Documentation of the BOF++ API is available at:

API reference documentation (Doxygen)