A Survey of Hybrid Braking System Control Methods

Abstract

With increasing attention being paid to electric vehicles, hybrid braking systems combining regenerative braking and conventional friction braking have become a hot research topic. Although some important advancements have been achieved in the field of hybrid braking system control, these have not been fully summarized. In order to fill this gap and provide a comprehensive perspective for other researchers, this paper surveys a wide range of research works reported in the literature on hybrid braking system control. We identify the advantages and limitations of existing hybrid brake control strategies via comparative analysis. Through analysis, we find that the control strategy used for brake torque distribution and braking systems’ coordinated control in current hybrid braking systems are usually designed separately. In order to ensure braking stability, most of the current hybrid braking control strategies are designed relatively conservatively, and it is difficult to fully leverage the advantages of hybrid braking systems. Comprehensively considering the coordinated control of braking torque distribution and braking systems is a good research direction for hybrid braking control research. Overall, this survey summarizes the existing research relevant to hybrid brake control methods and also discusses the research challenges and new research directions.

1. Introduction

Electric vehicles (EVs) have been the focus of the automotive industry in recent years, alongside growing concern regarding global environmental and energy issues [1]. Studies have shown that, in city driving situations, approximately between one-third and one-half of the energy of vehicles is consumed during deceleration [2]. Electric vehicles have the function of regenerative braking due to their use of electric motors. Regenerative braking is an effective method for improving the energy efficiency of electric vehicles by converting the vehicle’s kinetic energy into electric energy during braking procedures. Due to their regenerative braking function, electric vehicles’ driving range can be effectively improved and their energy utilization efficiency further enhanced. However, the braking torque of regenerative braking is limited and cannot meet the braking torque requirements of vehicles under all braking conditions. A friction braking system is still needed to guarantee the vehicle’s braking performance and safety. Thus, most EVs adopt hybrid braking systems [3].

Unlike conventional single braking systems, hybrid braking systems combine two braking systems with different characteristics. The hybrid braking system has the advantages of braking energy recovery and fast response speeds compared with conventional single braking systems. Usually, a hybrid braking system is composed of a regenerative braking system and a friction braking system. The friction braking system could be an electrohydraulic braking system (EHB), an electromechanical braking system (EMB), and so on [4]. However, different braking systems present quite different dynamic characteristics. Regenerative braking systems have fast response speeds. However, the maximum braking torque that can be provided by regenerative braking systems is relatively low, and it is difficult to meet the braking torque requirements in all braking situations. The EHB can meet the braking torque requirements under various braking conditions. However, the response speed of the EHB is slower than the response speed of the regenerative braking system. The difference in the dynamic characteristics of each braking system leads to a large braking jerk under the transitional condition of the hybrid braking system [5]. In addition, the regenerative braking energy recovery efficiency of the motor varies at different speeds and braking torques. An effective braking torque distribution strategy and a collaborative control algorithm can efficiently improve the vehicle’s braking comfort and the braking energy recovery efficiency of hybrid braking systems [6]. Thus, there are two key elements at work in the control of a hybrid braking system: first, the distribution of braking torque between different braking systems, and, second, the coordinated control strategy between different braking systems in the hybrid braking system.

Although there have been many studies on the distribution of braking torque and coordinated control in hybrid braking systems, analyses of and comparisons between these different methods are lacking. Currently, most review papers focus on the types of brake actuators, the development, control method, and application prospects of brake-by-wire actuators (such as electromechanical braking systems, eddy current–hydraulic systems, etc.), regenerative braking systems, and the energy recovery of regenerative braking systems [7,8,9,10,11,12,13,14]. However, there are no relevant review papers that comprehensively consider the braking torque distribution strategies and coordinated control strategies of hybrid braking systems. This paper aims to fill this gap in the literature by reviewing the braking torque distribution strategies and coordinated control strategies of hybrid braking systems used to improve the braking performance of hybrid braking systems.

The main contributions of this paper are to evaluate the current state of the field of hybrid braking system control, identify the main research challenges, and make recommendations for future research directions on hybrid braking systems.

The remainder of this paper is structured as follows: Section 1 provides a brief introduction to various forms of braking systems and hybrid braking systems. Section 2 discusses braking torque distribution strategies for hybrid braking systems and the coordination control methods of hybrid braking control. Section 3 presents the main research challenges, based on the discussion in the previous section. Finally, Section 4 offers conclusions regarding the current state of the field and provides recommendations for the direction of future research.

2. Braking Torque Distribution Strategy and Coordination Control of Hybrid Braking Control Methods

As shown in Figure 1, the braking process of a vehicle with a hybrid braking system mainly comprises braking intention, the braking command tracking control module, the braking torque distribution module, the brake system coordination module, and the hybrid brake system actor.

Braking intention typically represents control commands from the upper level of autonomous driving or braking commands from the driver. It is usually sent to the lower level in the form of a braking deceleration braking command signal ($a_x$). The braking command tracking control module is used to control the deceleration of the vehicle, tracking the upper braking command signal through a control algorithm. The braking command tracking control module outputs a total braking torque command signal ($T_{td}$) through the control algorithm. The braking torque distribution module distributes the braking torque to the brake system coordination module based on the total braking torque command of the upper layer and the characteristics of each braking system. The brake system coordination module is used to coordinate different brake systems to better track the total brake torque command signal. The hybrid braking system actuator module is the execution layer of the system and outputs braking torque to act on the vehicle. The braking command tracking control module, braking torque distribution module, and brake system coordination module are the three most important components of hybrid braking control [15].

The braking command tracking control module of the hybrid braking system is used to allow the vehicle to track the braking commands of the upper level. Many control algorithm for hybrid braking system have been proposed (As shown in

Table 1. A summary of control algorithm for hybrid braking system.

Refs.	Control Algorithm
[20]	Sliding mode control
[21,22]	Predictive control
[23,24]	Fuzzy control
[25]	Nonlinear control
[26]	PID
[27]	Neuro-fuzzy control
[28,29]	Robust control
[30]	Adaptive Neuro Fuzzy
[31]	Adaptive control

). The hybrid braking system has faster response speeds and greater braking torque than traditional single hydraulic braking systems. The hybrid braking system has effectively improved the performance of vehicles in many braking-related aspects, including vehicle stability [16], anti-lock braking systems [17,18,19], braking comfort [20], and yaw dynamics [21]. In order to allow vehicles to effectively track the upper-level brake command, a large number of control algorithms have been proposed. Sliding mode control is widely used in various industries due to its anti-interference and fast response capabilities, and it is also used in vehicle braking control. In [17], the braking performance of a hybrid electric braking system with a sliding mode controller is evaluated.

The overall goal of the braking torque distribution module and the braking coordination control module is to improve the efficiency of braking energy recovery while ensuring braking stability [32,33,34]. However, the brake torque distribution module and brake coordination control module each have different focuses. The braking torque distribution module focuses on recovering braking energy, while the braking coordination control module focuses on coordinating the dynamic response of different actuators. The braking torque distribution module and the braking torque coordination module have overlapping and complementary parts in terms of functionality. Therefore, in some controller designs, only the brake torque distribution module or the brake torque coordination module may appear; meanwhile, in other controller designs, both may appear. Usually, the braking torque distribution module is placed before the braking torque coordination module. In traditional vehicle braking torque distribution strategies, the braking torque distribution mainly considers the different vertical load distribution of the front and rear wheels caused by changes in the center of mass during the braking process. In order to fully utilize the braking force provided by the ground, the braking torque of the front and rear wheels is distributed along a specific curve based on the characteristics of the vehicle. This braking torque distribution strategy mainly targets the braking torque distribution during the longitudinal braking of vehicles. In addition to longitudinal braking, braking control during steering is also important; it directly affects the stability of the vehicle. The braking torque distribution of the four wheels needs to be coordinated to ensure the braking stability of the vehicle. A typical application is electronic stability control (ESC) [35,36,37,38]. In traditional vehicle braking control, the braking torque distribution method described above mainly focuses on the distribution of the braking force between different wheels and pays less attention to the braking torque distribution between different brake actuators.

With the development of new electric vehicles and due to the addition of motor regenerative braking, the distribution of the hybrid braking system’s braking torque not only considers the distribution of the braking torque between different wheels but also takes into account the distribution of the braking torque between different braking systems.

The braking torque distribution of a hybrid braking system is influenced by many factors, such as the external characteristics of the motor, state of charge (SoC) limitations, and load transfer [39,40,41,42,43]. In addition to these factors, the most important aspect of the braking torque distribution strategy is the distribution of braking torque based on braking strength. To ensure the vehicle’s braking stability, one widely used braking torque distribution strategy based on braking strength is that only the motor regenerative braking system works when the braking strength is lower than a certain set value, and only the hydraulic braking system works when the braking strength is higher than a certain set value or when the anti-lock brake system (ABS) is triggered. When the vehicle is operating under normal braking conditions, the braking torque distribution of hybrid braking systems based on braking strength can be roughly divided into two types: a specific proportional relationship or optimized proportion (as shown in Figure 2).

As shown in Figure 2, using a specific proportion means that, as the braking intensity increases, regenerative braking provides braking torque in a specific proportion. One specialized method for the proportional distribution of braking torque involves filtering the total braking torque demand (i.e., the braking command signal) and using the low-frequency part as the braking command signal of the friction braking system. The high-frequency part is used as the braking command signal of the motor regenerative braking system; then, the two are combined to output the braking torque to act on the vehicle. The above braking torque distribution strategy mainly considers the braking capacity of the motor, which is the maximum braking torque that the motor can provide. Unlike the specific proportional relationship method, the optimized proportional braking torque distribution method not only considers the braking capacity of the motor but also focuses on the braking energy recovery efficiency of the motor. As shown in Figure 2, the regenerative braking torque of the motor does not positively increase with the braking strength. The distribution of braking torque should consider the braking energy recovery efficiency of the motor at different braking speeds. When distributing braking torque, it is necessary to try to make the motor work in the range with the highest braking energy recovery efficiency to achieve the maximum braking energy recovery.

Table 2. Summary of braking torque.

Classification	Influence Factor	Control Algorithm	Advantages of Control Algorithms	Refs.
A specific proportional	Anti-lock braking system (ABS)	Nonlinear-sliding-mode-type control	Fast response speed	[17]
		Adaptive neuro fuzzy	Adaptability	[30]
		PID control	Model-free	[44]
		Fuzzy	Interpretability	[45]
	ESC	Fuzzy-rule-based control	Adaptability	[46]
		-	-	[47,48]
	All conditions	Adaptive fuzzy control	Adaptability	[49]
		-	-	[50]
	Safety-critical driving maneuvers	Sliding mode control	Fast response speed	[51]
	Road grade preview	-	-	[52]
	Two-wheel front driven	Sliding mode control	Fast response speed	[53]
	Load variation and wheel slip considerations	PID	Model-free	[26]
	-	Hybrid theory	Adaptability	[54]
	Different road surfaces	Fuzzy logic	Interpretability	[55]
	ECE regulations	-	-	[56]
	-	Predictive braking control	Optimization	[22]
Optimized proportion	ABS	Nonlinear Control	Accuracy	[25]
		Robust control	Robustness	[29]
		Self-organizing function-link Fuzzy cerebellar model Articulation controller	Adaptability	[57]
		Optimal control	Optimization	[58]
	ESC	LQR	Optimization	[31]
		Optimal control	Optimization	[59]
	All conditions	Model predictive control	Optimization	[56]
		Optimal control	Optimization	[56,60,61]
		Neural-network sliding mode control	Accuracy	[62]
		Fuzzy Logic Control	Interpretability	[63,64]
		Linear and nonlinear model predictive control	Optimization	[65]
	Battery	MPC	Optimization	[66,67]
	Braking strength	-	-	[68]
	Shaft vibration	Model predictive control	Optimization	[69]
	Sliding braking condition	Multi-objective optimization strategy	Optimization	[70]
	Road conditions and driver’s intentions	Model predictive control	Optimization	[71]
	Downshifting strategy	Stochastic dynamic programming	Optimization	[72]
	Driver’s braking intention	Model predictive control	Optimization	[73]
Other factors	Battery state	-	-	[74]
	Braking safety and ride comfort	-	-	[75]
	ECE R13H	Robust sliding mode controller	Robustness	[53]
	All conditions	Artificial neural network	Accuracy	[76]
		Energy constraint control (ECC)	Robustness	[77]
	Battery/supercapacitor	-	-	[78]
	DC/DC	Genetic algorithm (GA)-fuzzy control	Optimization	[79]

presents a summary of braking torque distribution strategies for hybrid braking systems.

Unlike the braking torque distribution module, which mainly considers the braking capacity (maximum braking torque) and braking energy recovery efficiency of each braking system, the braking torque coordination module mainly accounts for the dynamic characteristics of each braking system to ensure that the hybrid braking system follows the braking command signal and achieves braking smoothness. Usually, the braking torque coordination module is placed after the braking torque distribution module to correct the command signals assigned to each braking subsystem from the braking torque distribution module (as shown in Figure 3 and Figure 4). Usually, the braking torque coordination module is divided into two types: with a compensation module (as shown in Figure 3) and without a compensation module (as shown in Figure 4).

As shown in Figure 3, the braking torque coordination module without a compensation module can be roughly divided into two categories: closed-loop control that is mainly based on regenerative braking and closed-loop control that is mainly based on friction braking. The method of closed-loop control based on regenerative braking mainly aims to improve the tracking accuracy of the hybrid braking system to the braking command signal by utilizing the fast response speed of the motor braking. As shown in Figure 3a, the command braking torque of the friction braking system $T_f^{*}$ comes directly from the braking torque distribution module $T_{fd}^{*}$. Unlike the command braking torque of friction braking systems, which comes directly from the braking torque distribution module, the command braking torque of regenerative braking systems is the braking command from the braking torque distribution module $T_{md}^{*}$ plus the braking torque difference $T_{bc}$ between the command braking torque of the friction braking system $T_f^{*}$ and the output braking torque of the friction braking system $T_f$. The advantage of the above strategy is that it can ensure the tracking accuracy of the hybrid braking system with regard to the braking command signal when the braking command signal is not large or when the change rate of the braking command signal is not large. However, this method can only ensure the tracking accuracy of the brake command signal when the torque of the brake command signal is small and the change rate of the braking torque command signal is not fast. Unlike the above methods, the method of closed-loop control mainly based on friction braking utilizes the high-braking-torque characteristics of the friction braking system to ensure that the effects of the braking command signal can still be guaranteed, even under a high braking command signal. As shown in Figure 3b, the command braking torque of a regenerative braking system $T_m^{*}$ comes directly from the braking torque distribution module $T_{md}^{*}$. The command braking torque of a friction braking system is the braking command from the braking torque distribution module $T_{fd}^{*}$ plus the braking torque difference $T_{bc}$ between the command braking torque of the friction braking system $T_m^{*}$ and the output braking torque of the friction braking system $T_m$. By controlling the braking torque allocated to the motor’s regenerative braking system, the braking energy recovery efficiency of the hybrid braking system can be improved. However, due to the slow response speed of the friction braking system, its error relative to the braking command signal is relatively large.

In order to overcome the shortcomings of the methods described above, some researchers have proposed using the braking torque coordination module with a compensation module. The braking torque coordination module with a compensation module can also be divided into two types: closed-loop control that is mainly based on regenerative braking and closed-loop control that is mainly based on friction braking (as shown in Figure 4).

As shown in Figure 3 and Figure 4, the biggest difference between the braking torque coordination modules with a compensation module and without a compensation module is that the module with a compensation module has an additional compensation module. The following formula is used to describe the response speed of each braking system:

$${\displaystyle \frac{1}{\tau s+1}}$$

(1)

where $\tau$ is the time constant of the braking system, and the values vary for different braking systems. As shown in Figure 4a, the command braking torque of the friction braking system $T_f^{*}$ comes directly from the braking torque distribution module $T_{fd}^{*}$. Unlike the strategy of using a braking torque coordination module without a compensation module, the command braking torque of a regenerative braking system using the strategy of a braking torque coordination module with a compensation module is the braking command from the braking torque distribution module $T_{md}^{*}$ plus the braking torque difference $T_{bc}$ between the command braking torque of the friction braking system $T_f^{*}$ and the output braking torque of the compensation module $T_f$.

As shown in Figure 4b, the command braking torque of a regenerative braking system $T_m^{*}$ comes directly from the braking torque distribution module $T_{md}^{*}$. The command braking torque of a friction braking system using the strategy of a braking torque coordination module with a compensation module is the braking command from the braking torque distribution module $T_{fd}^{*}$ plus the braking torque difference $T_{bc}$ between the command braking torque of a regenerative braking system $T_m^{*}$ and the output braking torque of a compensation module $T_m$.

Table 3. Summary of braking torque coordination strategy.

Classification	Subclass	Control Algorithm	Advantages of Control Algorithms	Refs.
Braking torque coordination without compensation module	Closed-loop control mainly based on regenerative braking	-	-	[80]
		Fuzzy control	Interpretability	[81]
	Closed-loop control mainly based on friction braking	-	-	[82,83]
		Model predictive control	Optimization	[84]
		Optimal control	Optimization	[85]
		-	-	[86]
		Fuzzy logic	Interpretability	[87]
		Sliding mode control	Fast response speed	[88]
		-	-	[89,90,91]
Braking torque coordination with compensation module	Closed-loop control mainly based on regenerative braking	Input-constrained-based sliding mode control	Robustness and fast response speed	[92]
		Active disturbance rejection controller	Robustness	[93]
	Closed-loop control mainly based on friction braking	Sliding mode control	Fast response speed	[94]
		MPC	Optimization	[95]
		-	-	[96]
Other factors	Parallel mode	Model predictive control	Optimization	[97,98]
	Fixed ratio	Sliding mode control and fuzzy logic control	Fast response speed and accuracy	[99]
		-	-	[100]

presents a summary of braking torque coordination strategies for hybrid braking systems.

3. Primary Research Challenges in Hybrid Braking System Control

The existing braking torque distribution strategy mainly relies on the braking strength and the braking torque provided by each braking subsystem as the basis for braking torque distribution, without considering the control boundary (attraction domain) of each braking subsystem and the rate of change of the braking command signal. This results in a conservative distribution strategy and leads to the problem of a large braking impact. In terms of coordinated control, because most studies do not consider the control boundary of each braking subsystem, i.e., the attraction domain of the control system (or reachable set, indicating that the state trajectory originating from any point in this region tends to infinity over time and can mostly converge to the equilibrium point region), a dual closed-loop strategy based on motor compensation control can only be used to ensure braking stability, to the extent that most coordinated control strategies are relatively conservative, resulting in insufficient braking energy recovery. In addition, most studies divide the optimal distribution and coordinated control of braking torque in composite braking systems into an upper-layer braking torque distribution module and a lower-layer coordinated control module, which are designed separately (as shown in Section 2). This makes it difficult for hybrid braking systems to simultaneously obtain the high-precision tracking performance of braking command signals and highly efficient braking energy recovery. At present, there are two issues in the research on the optimal distribution and coordinated control of braking torque in composite braking systems: (1) Due to the failure to integrate with the braking torque distribution layer, and because the control boundaries (i.e., the attraction domain) of each braking subsystem are not considered during coordinated control design, the braking coordination control strategy developed to ensure braking stability is usually conservative, resulting in insufficient braking energy recovery, and (2) due to the failure to comprehensively consider the rate of change of braking command signals, the control boundaries of each braking subsystem (i.e., the attraction domain), as well as coordinated control strategies at the lower level when formulating the braking torque distribution strategy, the designed braking torque distribution strategy is also generally conservative. This ensures braking stability but results in limited braking energy recovery and relatively low braking impact. There is still a lack of research on improving braking smoothness and braking energy recovery while ensuring braking stability; such research should comprehensively consider the characteristics of braking command signals and the effects of various braking subsystems on braking stability, smoothness, and braking energy recovery. Moreover, brake torque distribution and coordinated control systems should be designed comprehensively rather than separately.

4. Conclusions

This paper presented a survey of hybrid braking system control. It was shown that research interest in this field has grown significantly in recent years. The research on hybrid braking system control is mainly divided into two categories: brake torque distribution and braking systems’ coordinated control. The methods discussed in this paper show great promise for the application of hybrid braking systems to vehicle control. However, the systems used for brake torque distribution and braking systems’ coordinated control in current hybrid braking systems are usually designed separately; thus, they cannot fully utilize the potential of hybrid braking systems, and there is significant room for improvement. In addition, although there are some designs that comprehensively consider brake torque distribution and braking systems’ coordinated control, they are often conservative in designing brake torque distribution and braking system coordinated control strategies due to the lack of control boundaries for each braking subsystem.

The main research challenges related to hybrid braking control were also discussed. As most studies do not consider the control boundary of each braking subsystem (i.e., the attraction domain of the control system), the braking torque distribution and coordination strategy of the hybrid braking system must be set relatively conservatively to ensure braking stability. This means that the braking energy recovery and other performance aspects of the hybrid braking system are conservative. Therefore, further research into the control boundaries of various braking subsystems is required. In order to comprehensively improve the braking energy recovery efficiency of the hybrid braking system while ensuring the tracking performance of the braking command in the hybrid braking system, further work is also needed to design a hybrid braking system with braking torque distribution and coordinated control based on the control boundaries of each braking subsystem.