The aim of the project is sonar based ground profile extraction for safe leg placement of walking machines. First development step is the construction of a sonar system consisting of four simultaneously operating ultrasonic range sensors.
One prerequisite to achieve this is each sensor's ability to identify the echo of its own transmitted pulse. This is only possible if the transmission pulse of each sensor differs significantly from the transmission pulses of all other sensors. That cannot be attained by a conventional sonar system, because all sensors operate with the same frequency. Hovever, if the sensors emit individual pseudo-random codes instead of conventional transmission pulses, each sensor is able to detect its own signal proportions in the received echo. Beyond that, each sensor is able to identify other sensors' signal proportions with the knowledge of their pseudo-random codes.
The use of such codes in connection with suitable signal processing algorithms gives substantial advantages compared to the use of conventional fixed frequency transmission pulses. Both, distance resolution and directional resolution can be improved significantly by triangulation. These are essential prerequisites for the extraction of ground profiles from ultrasonic data.
The project is promoted by the German Research Community (Deutsche Forschungs Gemeinschaft DFG ) and is part of the DFG programme Autonomous Walking