Effective and Fast Estimation for Multi-Source Navigation Sensor Reliability

As a new navigation technology, the idea of all source navigation is using any available sensor to achieve high-precision location service. However, navigation devices and data are susceptible to various environments and attacks, which emphasizes the need for pervasive security measures like data reliability evaluation on the fly. There exist several sensor data reliability measures which are based on the Internet of things devices, but these methods are weak for highly dynamic navigation data. This paper proposes a real-time navigation sensor data reliability evaluation method, which combined self-evaluation and mutual evaluation. Given the strong dynamics of navigation sensors, we improve the classical sensor anomaly judgment criteria and realize the data reliability judgment in high dynamic scenarios. Then, the spatial-temporal correlation and probability distribution of multi-sensor data are aggregated, and the aggregation results are combined with the self-evaluation reliability to realize the real-time evaluation of sensor data reliability. The performance of our method is evaluated using both trustworthy and untrustworthy data. The data of the former is collected by unmanned vehicles in different scenarios. The latter consists of two parts: the data simulated according to the anomaly model and the data collected when the actual sensor is blocked. Experiments demonstrate proposed method can effectively evaluate the reliability of sensors data.