A tightly coupled global navigation satellite system (GNSS)-Vision-IMU-wheel odometer (GVIWO) system is proposed, which can realize robust positioning in extreme environments. The purpose of this study is to achieve adaptive initialization in complex environments, sensor anomaly detection and processing, and adaptive robust localization in extreme environments.
Adaptive initialization includes traditional dynamic and static initialization and extreme condition initialization. To deal with the unstable visual features in the state of excited motion, a method of wheel odometer assisted initialization is designed. According to the abnormal condition of the sensor, the anomaly detection and attenuation mechanism are designed to realize the accurate positioning of the sensor under abnormal condition.
Tight coupling optimization of GNSS signals, RGB+Depth Map cameras, inertial measurement units and wheel odometers ensures accurate positioning in both indoor and outdoor environments. Through open data sets and field validation experiments, the proposed tightly coupled system has strong adaptability, especially in extreme environments.
A new framework is proposed by integrating GNSS, visual, inertial measurement unit (IMU) and wheel odometer sensors to form an efficient positioning solution. An adaptive initialization method is proposed to enhance the robustness and real-time performance of the positioning system in complex and dynamic environments. A mechanism for detecting and attenuating sensor anomalies is designed, enabling quasideterministic positioning under sensor anomalies.
