Shufeng Han
John Deere Intelligent Solutions Group

The GPS-based vehicle navigation technology has been successfully commercialized since the introduction of GPS for civilian use in the late 1980s. However, many agricultural vehicle navigation applications will require local perception sensors to either complement or substitute the GPS to meet the positioning requirements. This presentation starts with examples of such applications for which the perception-based navigation technologies are required. A number of perception sensors, including dead-reckoning, geo-magnetic direction sensors, ultrasonic sensors, vision, and scanning lasers, have been studied and applied to navigation for many years. Currently, vision and scanning lasers are the most promising sensors for perception-based navigation. Major data processing and analysis algorithms for these two types of sensors are presented. Implementation examples using vision and scanning lasers for automated guidance of agricultural tractors are given. In contrast to GPS-based navigation technology, there are still significant challenges for commercialization of the perception-based navigation technologies. The current trend is to use sensor-fusion-based approach, i.e., the combination of GPS and perception sensors, to achieve better positioning accuracy and robustness.