UNMANNED SYSTEM FOR ACQUISITION OF GEOSPATIAL DATA

Abstract

In recent years, GNSS technology along with miniaturized MEMS sensors creates the possibilities of the development of low-cost unmanned aerial system for acquisition of geospatial data. Such systems are mostly adopted for inaccessible areas due to their low-cost and high-resolution data. However, the acquired data require the control points to improve the accuracy. But the control points establishment in inaccessible areas is costly, time consuming and risky. Thus, the objective of this research work is to develop an unmanned ground system for acquisition of geospatial data by making use low-cost sensors and its outcome will be beneficial for a surveyor to study inaccessible areas. Also, it will reduce overall cost and time consumption for data acquisition in hazardous environment. The unmanned ground system for acquisition of geospatial data is a new concept along with aerial and satellite system. The non-vision and vision sensors have been used to develop an unmanned ground system. As vision sensors are not feasible during night, non-vision sensors (such as GNSS, INS and wheel odometry) are preferred to develop unmanned systems. In previous works, in absence of GNSS measurements the navigation solution has been computed using INS and Odometry sensors. However, the heading information provided by INS unit depends on magnetometer measurements. The major issue with magnetometer sensor is its sensitivity about ferrous objects present in surrounding environment. Therefore, need to develop a navigation algorithm for low-cost sensors to address the mention limitation. In this research, the proposed concept is implemented in four stages: starting with sensor calibration and testing, followed by data fusion, system field testing and finally data acquisition using low-cost GNSS receiver and performance evaluation. The calibration process is used to minimize or entirely eliminates the errors presents in the sensor measurements. After this, the performance of individual sensor has been tested at institute laboratory. The second stage of system development is multi-sensor i integration using data fusion technique. The Kalman Filter has been used for data fusion. Data fusion improves the reliability and eliminates the limitations associated with individual sensors. Also, an algorithm has been developed to eradicate the impact of ferrous objects on navigation solution. This has been achieved by integration of gyro and odometry sensor while ignoring magnetometer measurement. In third stage, the field testing has been conducted of the developed system. Finally, an inexpensive RTK GNSS receiver has been mounted on the system and data acquisition process has been conducted for selected study areas. The Stop-and-Go method has been adopted for acquisition of data. In order to determine the reliability of the proposed navigation algorithm ferrous objects have been kept in the path. From the obtained results, it has been noticed that system can be provide a reliable navigation solution under ferrous objects up to 60 m. During calibration it has been noticed that the maximum impact of a Ferrite ring magnet with magnetic flux density of 0.35 T is up to 2 m. Hence, 60 m navigation from the point of first ferrous objects detection is enough for avoid impact of surrounding ferrous objects. The qualitative evaluation of the system geospatial data is done by comparing the reference data. The reference data of the study areas have been collected using geodetic GNSS receiver. It has been noticed that system GNSS receiver is capable to provide the data with accuracy of 3 cm in real-time.