Tightly Coupled RTK/MIMU using Single Frequency BDS/GPS/QZSS Receiver for Automatic Driving Vehicle
Qifen Li, Beijing Research Institute of Telemetry, China; Li-Ta Hsu, Hong Kong Polytechnic University, Hong Kong; Yanlei Gu, Shunsuke Kamijo, Institute of Industrial Science, The University of Tokyo, Japan
Autonomous driving is a main research focus in automobile industries. Recently, Google, along with several automobile manufacturers has rapid progress in this field, leading them to test it on public roads. Precise positioning technology is one of the keys in this advance technology. It is relatively easy to obtain its precise position (i.e., 10 centimeters) using real time kinematic (RTK) in open sky area. Considering the vehicle will be operated in all kinds of environment, the positioning system should be able to provide robust localization information in harsh environment. Most of the driving environment is blocked by trees or high buildings, it is especially difficult to get a “fixed” solution for RTK algorithm in such environments. Sometimes, it is also difficult to get position information by single point positioning (SPP) because less than four satellite are visible. This paper describes a tightly-couple integrated positioning system using MEMS IMU and low-cost single frequency GNSS receiver for vehicle driving in urban area. The design objective is to realize accuracy and robust position in various streets environment.
In Tokyo, the observation conditions of GPS and BDS is perfect. At the same time, Japan has their own regional navigation system called QZSS, which is designed to stay high elevation for Japanese region. Thus, BDS/GPS/QZSS system is selected in this paper. In fact, there are several frequency bands in each navigation system. BDS have B1, B2, B3 bands, GPS have L1, L2, L5 bands, and QZSS is similar to GPS. Accordingly, there are several kinds of GNSS receivers, single frequency, dual frequency or triple frequency receivers. The advantage of the dual frequency receiver is to eliminate the effect of ionosphere. However, it is more expensive than single receiver. Meanwhile, with the rapid development of the integration circuit technology, the channel capability of satellite navigation receiver increases quickly, resulting in the price of multi-constellation system receiver is close to the price of single system receiver. In order to increase the number of the visible satellite in harsh environment, this paper applies a BDS/GPS/QZSS receiver to provide the raw observations. But it is still a problem for vehicle to drive in tunnel because none of the satellites are visible.
Inertial navigation system (INS) consists of three axis accelerometers and three axis gyroscopes, which is generally called inertial measurement unit (IMU). After initialization, INS is applied to provide continuously navigation service while the GNSS solution is not available. However, if there is no correction, it will gradually diverge because of error accumulation. Comparing with GNSS solution, IMU is short-term high accuracy but error accumulating. In the market, there are three levels of IMU. Navigation grade IMU is the most precise but the most expensive equipment. At the same time, the volume and power consumption is huge for automobile applications. Tactical level one is the second precise equipment. It is still high-cost for autonomous driving market. The consumer grade IMU is usually made by MEMS technology. The power consumption and the price of MEMS are reasonable. Although MEMS IMU (MIMU) is still low precision now, the accuracy is gradually increasing recently and its price, volume and power consumption have undefeated features for consumer market, this paper select MEMS IMU as the test equipment.
Finally, this paper applies coupled information to correct the error of MIMU and to detect and recovery multipath and cycle slip. These raw measurements can be pre-processed before applying RTK algorithm. By reducing the environment influence to the pseudorange and carrierphase, the fix ratio or the accuracy of the float solution can be improved. The technology mentioned in this paper can be used to develop a low-cost precision position system for automatic driving vehicle. The detail of the methodology and experiment setup can be found in the attachment.
RTK, IMU, Multipath, cycle slip