A Multi-Element System of Surrounding Recognition And Objects' Localization for Unmanned Ground Vehicles

Abstract

Equipping engineering machines and trajectory vehicles with automatic and remote control systems is necessary in case there is any environmental hazard for an operator i.e. environment conditions, an impact from fire field etc. It is particularly crucial when there is no possibility for any person to be in a machine or in its close surrounding. It applies to both extreme environment conditions (high temperature, pressure or environment contamination) and the likeliness of direct man health and life hazard (i.e. removal, disposal or neutralization of hazardous materials or area demining -particularly when the enemy operates on a fire field). In all above cases of using engineering machines and trajectory vehicles there is a need to remote control without the possibility of using direct impulse feedback by operators. That is the reason why it is so necessary to work on elaboration vision system enabling determination of the machine location in geodesic system and operating accessories configuration. This paper describes an overall characteristics of the steering unit in an unmanned ground vehicle depending on tasks to be performed, considering the drive and the steering's system structure required. Current development of visual systems used in unmanned ground vehicles has been shown. A visual system to be used in remote controlled machines and ground vehicles has been presented. Its structure and limitations within picture depth estimation resulting from the system's structure have been presented. Furthermore the system of defining the cameras' location in an external reference system has been described.

Use of Laser Rangefinder to Detecting in Surroundings of Mobile Robot the Obstacles

“…3). The resulting range measurement lies somewhere between the distance to the object and the distance to the background [6]. In researching the mixed pixels phenomenon we found that if the distance AD between the edge of the foreground object and the background is close to the laser pulse width, AL (AL » 1 m for the Sick LRF), then a substantial number of mixed pixels is generated.…”