Machine Learning-Based Prediction Models and Threat Detection for Lane-Keeping Assistance
Doctoral thesis, 2023
Traffic accidents have been an ongoing problem for over a century and many efforts have been made to improve traffic safety. Historically, the focus has been on passive safety with innovations, such as crumpling zones, three-point seat belts, and airbags, that aim to mitigate the impact of collisions. As technology advanced, the focus shifted toward active safety, which aims to avoid accidents.
Advanced driver assistance systems are nowadays utilized in vehicles to support the driver in critical situations where the driver is likely to fail the driving task. The system uses sensor information to estimate the risk of a threatful event, such as an unintended lane departure, and decides whether an automatic avoidance maneuver should be activated. However, from a legal perspective, it is the driver who is responsible for the driving, and consequently, the driver must be able to override an erroneous maneuver. This is an important aspect, as it restricts the system to the use of low-intensity maneuvers. That implies that a maneuver needs to be activated sufficiently early in time to be able to avoid the threatful situation, i.e., a long prediction horizon is needed to detect the threat in time.
The decision to intervene with a supportive automatic avoidance maneuver is based on the output from the threat assessment, which uses a prediction model to estimate how the current traffic situation is evolving with time. Designing a well-functioning prediction model is challenging, as it must deal with multiple sources of uncertainties, such as sensor noise and drivers’ intentions, and it becomes even harder as the prediction horizon increases.
This thesis focuses on how machine learning can be used to improve the performance of a lane-keeping assistance system. The goal has been to develop learning-based prediction models that are high-performing, robust, and efficient to compute in real time. The approach has been to evaluate the performance of linear and non-linear regression models using real-world data. The results show that both linear and non-linear prediction models are significantly better than a kinematic model. It also shows that linear prediction models are nearly as good as non-linear models, especially for shorter prediction horizons. However, the linear model is significantly easier to compute in real time and may therefore be a sufficient alternative for applications where computational power is restricted. Moreover, the robustness towards anomalies and samples that are out of the operational design domain can be improved by utilizing uncertainty-aware prediction models.
Advanced driver assistance systems are nowadays utilized in vehicles to support the driver in critical situations where the driver is likely to fail the driving task. The system uses sensor information to estimate the risk of a threatful event, such as an unintended lane departure, and decides whether an automatic avoidance maneuver should be activated. However, from a legal perspective, it is the driver who is responsible for the driving, and consequently, the driver must be able to override an erroneous maneuver. This is an important aspect, as it restricts the system to the use of low-intensity maneuvers. That implies that a maneuver needs to be activated sufficiently early in time to be able to avoid the threatful situation, i.e., a long prediction horizon is needed to detect the threat in time.
The decision to intervene with a supportive automatic avoidance maneuver is based on the output from the threat assessment, which uses a prediction model to estimate how the current traffic situation is evolving with time. Designing a well-functioning prediction model is challenging, as it must deal with multiple sources of uncertainties, such as sensor noise and drivers’ intentions, and it becomes even harder as the prediction horizon increases.
This thesis focuses on how machine learning can be used to improve the performance of a lane-keeping assistance system. The goal has been to develop learning-based prediction models that are high-performing, robust, and efficient to compute in real time. The approach has been to evaluate the performance of linear and non-linear regression models using real-world data. The results show that both linear and non-linear prediction models are significantly better than a kinematic model. It also shows that linear prediction models are nearly as good as non-linear models, especially for shorter prediction horizons. However, the linear model is significantly easier to compute in real time and may therefore be a sufficient alternative for applications where computational power is restricted. Moreover, the robustness towards anomalies and samples that are out of the operational design domain can be improved by utilizing uncertainty-aware prediction models.
Machine learning
ADAS
Threat detection
Threat assessment
Room HC4, Hörsalsvägen 14
Opponent: Professor Pongsathorn Raksincharoensak,Tokyo University of Agriculture and Technology
Author
John Dahl
Chalmers, Electrical Engineering, Systems and control
Advanced driver assistance systems (ADAS) have become increasingly important in modern vehicles, offering a range of features that enhance safety, convenience, and comfort for drivers. Technologies such as lane departure warnings, automatic emergency braking, and adaptive cruise control, have significantly reduced the risk of accidents, injuries, and fatalities. However, recent utility analysis has shown that there is still room for improvement in the ADAS’s effectiveness, where the main challenges are to include the driver’s intention in the system design and improve the system’s robustness towards anomalies and disturbing noise in the sensor data.
An ADAS uses an array of sensors, such as cameras, lidar, radar, and driver monitoring sensors, to estimate the current traffic situation. This information is used in the threat assessment, which is an algorithm that predicts how the traffic situation will evolve over time. The threat level is calculated based on the prediction using threat metrics, which are typically tailored to indicate the risk of a certain event. For example, the time-to-lane crossing metric computes the time until the vehicle crosses a lane marker, where a low metric value can be interpreted as a high risk of departing from the lane. The threat metrics are used in the decision-making to detect whether the driver should be supported by an automatic intervention, or not. If the vehicle is at risk, an intervention trajectory is planned, which is followed using the mechatronic actuators of the vehicle.
The focus of this thesis is on how to improve the performance of a lane-keeping assistance system using a machine-learning approach. Specifically, the aim has been to develop efficient threat assessment and decision-making algorithms that can include the driver’s intention, and handle anomalies and noise in the input data to achieve a high system performance and robustness. The experiments and analysis are based on real-world data.
An ADAS uses an array of sensors, such as cameras, lidar, radar, and driver monitoring sensors, to estimate the current traffic situation. This information is used in the threat assessment, which is an algorithm that predicts how the traffic situation will evolve over time. The threat level is calculated based on the prediction using threat metrics, which are typically tailored to indicate the risk of a certain event. For example, the time-to-lane crossing metric computes the time until the vehicle crosses a lane marker, where a low metric value can be interpreted as a high risk of departing from the lane. The threat metrics are used in the decision-making to detect whether the driver should be supported by an automatic intervention, or not. If the vehicle is at risk, an intervention trajectory is planned, which is followed using the mechatronic actuators of the vehicle.
The focus of this thesis is on how to improve the performance of a lane-keeping assistance system using a machine-learning approach. Specifically, the aim has been to develop efficient threat assessment and decision-making algorithms that can include the driver’s intention, and handle anomalies and noise in the input data to achieve a high system performance and robustness. The experiments and analysis are based on real-world data.
Areas of Advance
Transport
Subject Categories
Electrical Engineering, Electronic Engineering, Information Engineering
ISBN
978-91-7905-834-0
Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5300
Publisher
Chalmers
Room HC4, Hörsalsvägen 14
Opponent: Professor Pongsathorn Raksincharoensak,Tokyo University of Agriculture and Technology