Under way
Research and Development of Localization System for a Vehicle involving Fusion of Data from Location Sensors
Type of work:
Master thesis
Assignment:
- Literature review: Basic research about fusion algorithms, GNSS systems and sensors, such as accelerometer, gyroscope, compass and GPS.
- Development: a) Application to forward position signals over Wi-Fi into data processing environment & b) Fusion algorithm
- Evaluation: a) Comparison between different smartphone & b) Comparison of fusion output against reference system
- Documentation: Writing thesis report consisting experimental setup, procedures, model diagram, results and conclusion
Skills:
- Basic understanding of GNSS systems / Smartphones and algorithms
- C++ / Scripting
- Wi-Fi / Network protocols
Background:
Accurate localization is one of the fundamental requirements for autonomous vehicles. This thesis aims to utilize multiple localization devices in order to improve the positioning accuracy of a vehicle. Firstly, multiple devices increase redundancy, which can be particularly useful in urban environments, in which buildings and other obstacles can block GPS signals.
Secondly, multiple devices allow us to carry out sensor fusion, which will reduce errors caused by outliers, sensor drift, mechanical and electrical faults etc. Moreover, additional information, such as vehicle actuators and the relative positioning of the localization devices mounted on the vehicle, could potentially further improve the overall accuracy.
The idea is to mount the multiple positioning devices, such as smartphones, on the vehicle and obtain GPS, accelerometer, gyroscope, compass data. A mobile application needs to be developed to forward this sensor information for further processing. This can involve steps like data synchronization and sensor fusion. This thesis would involve a literature review of the Kalman filter algorithm, which has been used for sensor data fusion in navigation applications since the 1960s. Possible variants as well as adaptations of the algorithm would be carried out for the use-case detailed in this thesis.
The position values of this new system are compared with the position values obtained from unfused sensor data. In case a reference system is available, unfused and fused data can be evaluated against it. The resulting data is visualized for different time steps during the localization process. Assessment on whether newly developed system with multiple smartphone GPS data fusion have positioning errors below acceptance limits can be performed.
Supervisor:
Student:
Akshay Laxman Powar
Year:
2023