Observation and modeling of mobile environment

Perception of the environment in a mobile robot is a difficult task. Other moving objects and various changes in the environment and motion of the robot must be handled together. Control of mobile movements in a known or especially in an unknown environment involves knowledge of exact details of other objects in surrounding areas. These details may include information such as instant position, speed and direction of movement, dimensions and collision possibility. The observation of environment is done by using various sensors, which can give information about distances and directions of other objects, which can be either stationary or moving. Depending on the accuracy demands of the application the number and complexity of sensors may vary a lot. Sensor signals are preprocessed to minimize the amount of data, that is transferred ahead to main processing units. To make this processing simple and easy data from different sensors must be presented similarly. For each transducer type there is an own filter to produce a common data representation. This representation is done in such a way that adding and removing of sensors is possible without modifying other parts of the system. The modeling of the mobile environment is based on the combined information from various sources. These sources include a priori data, e.g., maps of the area where mobiles are used, measured distances, and data of other objects. This environment model could be described as a dynamic map of the mobile environment.