A 10,000-word article explains the current status and trends of the development of commercial vehicle electronic steering systems

In addition, the research on chassis domain integrated control based on wire-controlled steering for passenger cars is becoming increasingly mature, while the research on chassis domain integrated control in the commercial vehicle field is still insufficient.

2.3 Assisted Driving and Autonomous Driving

At present, due to the incomplete laws and regulations on autonomous driving and the insufficient security of related algorithms, it is difficult to apply autonomous driving algorithms for passenger cars on urban roads. Unlike passenger cars, which mostly travel on urban roads, a considerable number of commercial vehicles travel in closed, unmanned and fixed-path conditions, such as mining areas, terminal freight, warehouse storage, structured highways, etc. In the above relatively simple or closed scenarios, commercial vehicle advanced driver assistance and autonomous driving will be easier to implement [68].

Thanks to the application of assisted driving functions, the number of collision accidents caused by commercial vehicles has decreased significantly in the past few years [69]. At present, research at home and abroad is still focused on L2 level functions, such as lane keeping assist (LKA), adaptive cruise control, lane departure warning, etc. These functions can help reduce vehicle collision accidents caused by driver distraction [70].

Unlike passenger cars, the EHCPS configuration of commercial vehicles has problems such as untimely actuator response and discontinuous steering wheel torque due to the inherent nonlinearity and hysteresis of the hydraulic subsystem. Reference [71] superimposed the target torque obtained by the anti-disturbance control and the target steering wheel angle calculated by the human-machine-in-the-loop MPC, and designed an LKA control strategy to overcome the influence of uncertain factors such as the perturbation of internal parameters of the steering system; Reference [72] adopted a variable weight multi-point predictive controller to suppress the nonlinearity and interference problems of the electro-hydraulic composite steering system in LKA control; Reference [73] designed an LKA control strategy based on the establishment of an electro-hydraulic steering system model and taking into account the driving characteristics of the driver.

In order to solve the problem of human-machine control rights allocation caused by human-machine collaboration in commercial vehicle assisted driving, the reference [71] studied the intervention of the LKA system on the driver, introduced the concept of intervention coefficient, and reduced the intervention on the driver while ensuring safe driving; the reference [74] solved the conflict problem when switching between the power correction mode and the driver mode in the LKA strategy. The reference [75] adopted adaptive MPC, combined the driver's cognition, muscle mechanics model and prediction model, and dynamically allocated the control authority, which reduced the conflict between the driver and the vehicle, and also improved the path tracking performance.

In the field of commercial vehicle assisted driving, the discontinuous steering wheel torque caused by the hydraulic system and the failure to consider the driver's driving characteristics result in a poor human-machine experience. At the same time, the human-machine conflict problem is still significant. How to ensure that the driver does not distrust the assisted driving system after being strongly interfered by other systems or other vehicles needs to be solved.

In the field of commercial vehicle autonomous driving, the control of truck platooning has been a hot topic of research in recent years. That is, a manually driven truck (the lead truck) is followed by multiple autonomous trucks. Reference [76] proposes a truck platoon coordination method based on a consensus algorithm to solve the problem of platoon formation and modification. The method enables trucks to exchange information about their current status in real time and adjust the platoon order and following distance based on the real-time information. Reference [77] studies the fuel saving problem of platoon control on a given route. Reference [78] further studies the speed planning and tracking control problems of platoon driving while considering fuel saving. Reference [79] considers the platoon control of heterogeneous vehicles, that is, the platoon driving control of vehicles of different types.

Commercial vehicle platoons usually travel on structured highways with relatively simple road conditions. There are other commercial vehicles or passenger cars interacting with the platoon. The interaction between the platoon and other vehicles needs further study.

In closed scenarios such as mining areas and docks, autonomous driving of commercial vehicles has a wide range of applications. References [80] to [82] studied autonomous driving in mining areas to ensure that mining trucks can reach their destinations with high precision in different scenarios. References [83] to [84] studied autonomous driving of commercial vehicles in port scenarios.

3 Functional safety requirements and fault-tolerant control strategies

3.1 Functional safety requirements

As automobiles have evolved from mechanical systems to complex systems that combine electrical, electronic and mechanical systems, safety and reliability have always been the focus of attention in the industry. With the development of electrical and electronic technology, the complexity of automobile steering systems has continued to increase, and the risks of system failures and random hardware failures have also increased.［85］

To address the above challenges, the ISO 26262 [86] standard, which defines functional safety for the automotive industry, was developed to ensure the safety of the overall system by eliminating unreasonable design risks or improving failure measures for mechanical failures, meaning that there is no unreasonable risk in operation when a failure occurs.

For commercial vehicle steering systems, vehicle-level hazards often include three types: vehicle lateral movement that does not follow or exceeds the driver's intention, insufficient lateral movement response, and loss of lateral movement control capability [87]. Based on the above three types of hazards, three functional safety goals corresponding to the steering system are proposed, namely, preventing automatic steering that violates the driver's intention, providing correct steering assistance, and maintaining steering control capability [88]. System-level faults generally include the following six types, namely, steering lock, steering failure, unexpected assistance, excessive steering assistance, insufficient steering assistance, and reverse steering assistance. The corresponding fault manifestations and vehicle-level hazards are shown in Table 3.

According to the system functional safety design and management process given by the ISO 26262 standard, the functional safety hazard analysis and risk assessment of the commercial vehicle steering system are carried out to determine the safety goals of the steering system. Based on the safety goals, safety analysis is carried out and corresponding fault diagnosis and fault-tolerant control strategies are formulated to ensure that the system failure risk is within an acceptable range when a system component (such as a sensor, actuator or controller) fails.

3.2 Fault diagnosis method

Commercial vehicle steering system faults mainly include actuator faults, sensor faults, communication faults, etc. [89], and the commonly used fault diagnosis methods are summarized in Figure 8.

The qualitative diagnostic method is mainly completed by hardware self-test, such as overvoltage, undervoltage, communication interruption, CAN bus short circuit and other faults. Then, the diagnostic logic is constructed based on engineering experience to make a qualitative judgment on the system fault.

Quantitative analysis methods include data-driven detection methods [92] and model-based detection methods [93]. Data-driven fault detection methods are driven by computing power and rely on a large amount of fault data to diagnose the current system status, such as neural networks [94] and deep learning [95]. They can quickly and accurately diagnose and locate faults, but when the data has a lot of noise, the confidence of the diagnosis will be greatly reduced.

Model-based fault diagnosis algorithms require a deep understanding of the physical or mathematical model of the system. Using the constructed system model and measurable information, observers or filters are designed. The residuals are calculated based on the observed output and the actual output of the system, and the residuals are then analyzed to perform fault diagnosis and fault location. Common system state observers include: Romberg observer [96], sliding mode observer [97], robust observer [98], etc.

In actual engineering applications, simple qualitative analysis combined with model-based fault diagnosis methods are often used to comprehensively diagnose system faults and locate faults [91].

3.3 Fault-tolerant control strategy

The fault-tolerant control of commercial vehicle steering systems mainly targets two types of devices: sensors and actuators. Sensor failure generally refers to inaccurate measurement values ​​caused by damage to sensor hardware, aging, etc. [99]; steering actuators are terminal devices of commercial vehicle steering systems. Mechanical damage such as wear and tear caused by frequent operation and environmental corrosion may cause system parameter perturbations, resulting in unexpected deviations between system instructions and execution actions [100].

There are two main methods of fault-tolerant control, namely active fault-tolerant control and passive fault-tolerant control. Passive fault-tolerant control mainly targets known possible faults and designs highly robust fault-tolerant control strategies to make the system insensitive to specific faults, thereby maintaining the stability of the system. The classic control block diagram is shown in Figure 9.