The pursuit of ultimate dynamic performance in road vehicles has led to the emergence of innovative chassis architectures. Among these, the Corner Module Architecture for electric vehicle cars represents a paradigm shift. By integrating an in-wheel motor drive system, a steer-by-wire system, a brake-by-wire system, and an active suspension system into a single, compact unit at each wheel corner, this architecture unlocks unprecedented potential for independent control of vehicle motion. This distributed and decoupled control capability makes the corner-module-based electric vehicle car an ideal platform for researching and implementing advanced stability and fault-tolerant control strategies. Each wheel can be controlled precisely in terms of driving/braking torque, steering angle, and vertical force, allowing the vehicle to achieve superior maneuverability, trafficability, and handling stability compared to conventional designs.
However, this high level of integration and performance comes with inherent reliability challenges. The compact packaging within each corner module can lead to thermal management issues for electronic components, increasing the risk of faults in the steering or drive systems. Furthermore, the increased unsprung mass associated with in-wheel motors can exacerbate the dynamic impact from rough road surfaces on these sensitive systems. The independent drive nature of this electric vehicle car can also lead to uneven tire wear, and the vulnerable position of the tires makes them susceptible to damage from road debris or sharp objects. A critical failure scenario is when a single wheel’s steering/drive system seizes (locks up) or when the tire sustains severe damage (e.g., puncture, blowout). In such cases, the conventional fault-tolerant strategies—which typically rely on redistributing steering angles or torques among the remaining functional wheels—become ineffective or even dangerous. The locked or damaged wheel creates significant and unpredictable longitudinal and lateral forces, posing a severe threat to stability and causing rapid tire degradation.
To address this challenging scenario, this paper proposes a novel chassis-coordinated fault-tolerant control strategy based on a three-wheel driving mode. The core idea is to utilize the active suspension system—assuming it remains operational—to physically lift the faulty wheel off the ground. The vehicle’s weight is then supported and propelled solely by the remaining three wheels. This approach fundamentally removes the disruptive influence of the faulty wheel. The primary challenge shifts from torque/angle redistribution to maintaining the overall stability of a three-wheeled electric vehicle car during motion, especially under acceleration, deceleration, and cornering. This requires a deep understanding of the stability boundaries for three-wheel support and a coordinated control framework that manages suspension height, wheel slip, and path tracking simultaneously.
This article is structured as follows: First, we conduct a theoretical analysis to elucidate the dynamic stability mechanism of the three-wheel driving mode. We derive the stable region for the vehicle’s center of gravity (CoG) under the physical constraints of suspension travel. Second, by combining this stability region with planned path information (curvature) and vehicle parameters, we calculate the maximum allowable longitudinal speed and lateral acceleration to ensure safe operation. Third, we design a hierarchical, chassis-coordinated fault-tolerant controller. The core of this controller is an active suspension controller based on Sliding Mode Control (SMC), responsible for maintaining the three-wheel stance. Supporting this core are two optimized ancillary controllers: a Traction Control System (TCS) based on Predictive Sliding Mode Control (PSMC) to suppress wheel slip induced by suspension actions, and a path-following controller based on Model Predictive Control (MPC) to ensure the electric vehicle car accurately tracks its intended trajectory. Finally, the effectiveness of the proposed strategy is validated through co-simulation using MATLAB/Simulink and CarSim.
1. Theoretical Analysis of Three-Wheel Driving Stability for the Electric Vehicle Car
1.1 Stability Mechanism and Center-of-Gravity Stable Region
Consider a four-wheeled electric vehicle car. When supported by all four wheels, the vertical forces from the sprung mass acting on each suspension sum to the total vehicle weight:
$$
\sum_{i=1}^{4} F_{zvi} = m_s g
$$
where $m_s$ is the sprung mass, $g$ is gravitational acceleration, and $F_{zvi}$ is the vertical force on suspension $i$ (i=1: Front-Left, 2: Front-Right, 3: Rear-Left, 4: Rear-Right). The force distribution depends on the CoG position. Let the initial CoG position be at point $O(0,0)$ in a coordinate system centered on the vehicle, with the x-axis pointing forward and the y-axis pointing left. The coordinates of the four suspension connection points (at the wheels) are: $1(L/2, B/2)$, $2(L/2, -B/2)$, $3(-L/2, B/2)$, $4(-L/2, -B/2)$, where $L$ is the wheelbase and $B$ is the track width.
Assume the faulty wheel is the Rear-Right wheel (wheel 4). The goal is to lift this wheel, making the electric vehicle car run on wheels 1, 2, and 3. For stable three-point support, the vehicle’s CoG must project inside the triangle formed by the contact patches of the three supporting wheels (wheels 1, 2, and 3). This triangle’s vertices are points 1, 2, and 3. The line connecting points 2 and 3 forms one edge of this triangle. A necessary condition for stability is that the portion of the sprung mass on the side of the lifted wheel (near point 4) is less than or equal to the portion on the opposite side, relative to the support triangle. A static analysis shows that stability requires the area of the rectangle associated with the lifted corner to be smaller than or equal to the area of the rectangle diagonally opposite.
If the CoG moves to a new point $O_t(x, y)$ due to body attitude changes (roll and pitch), stability is maintained if the following inequality holds relative to the support triangle formed by wheels 1,2,3:
$$
\left( \frac{L}{2} – x \right)\left( \frac{B}{2} – y \right) \leq \left( \frac{L}{2} + x \right)\left( \frac{B}{2} + y \right)
$$
Solving this inequality defines the stable region as the area above the line:
$$
y \geq -\frac{B}{L} x
$$
This region corresponds to the area to the “upper-right” of the line connecting support points 2 and 3, ensuring the CoG is within the support triangle. However, this theoretical stable region is further constrained by the limited travel of the active suspensions. The body’s pitch and roll angles ($\theta$ and $\phi$) are limited by the maximum extension/compression of the suspensions. Given the distances from the CoG to the pitch center ($h_p$) and roll center ($h_r$), the physical displacement limits of the CoG due to suspension travel are:
$$
\begin{aligned}
x_{\min} &\leq x \leq x_{\max}, \quad \text{where } x_{\min} = -h_p \sin(\theta_{\min}), \ x_{\max} = h_p \sin(\theta_{\max}) \\
y_{\min} &\leq y \leq y_{\max}, \quad \text{where } y_{\min} = -h_r \sin(\phi_{\min}), \ y_{\max} = h_r \sin(\phi_{\max})
\end{aligned}
$$
where $\theta_{\min}, \theta_{\max}, \phi_{\min}, \phi_{\max}$ are the minimum and maximum achievable pitch and roll angles. The intersection of this rectangular region (from suspension limits) and the theoretical stable half-plane $y \geq -(B/L)x$ forms the final, physically achievable CoG Stable Region for the three-wheeled electric vehicle car, typically a trapezoidal or polygonal shape. For a typical electric vehicle car with $L=2.6$ m, $B=1.4$ m, $h_p = h_r = 0.4$ m, and max pitch/roll angles of $\pm 6.6^\circ$ and $\pm 12.4^\circ$ respectively, the stable region boundaries are:
| Boundary | Equation / Constraint | Physical Meaning |
|---|---|---|
| Left (Min Pitch) | $x \geq -46.2 \text{ mm}$ | CoG cannot shift too far rearward. |
| Right (Max Pitch) | $x \leq 46.2 \text{ mm}$ | CoG cannot shift too far forward. |
| Bottom (Min Roll) | $y \geq -85.7 \text{ mm}$ | CoG cannot shift too far right. |
| Top (Max Roll) | $y \leq 85.7 \text{ mm}$ | CoG cannot shift too far left. |
| Theoretical Stability | $y \geq -0.538x$ | CoG must be inside support triangle. |
1.2 Dynamic Constraints: Longitudinal and Lateral Acceleration Limits
When the three-wheeled electric vehicle car is in motion, inertial forces due to acceleration and cornering cause load transfer, which shifts the effective CoG position. To maintain stability, this dynamic CoG shift must remain within the static CoG stable region derived above.
Longitudinal Acceleration Limit (Pure Acceleration/Deceleration): During straight-line acceleration, a pitch moment $m_s a_x h_p$ is induced, where $a_x$ is the longitudinal acceleration. This is countered by the restoring moment from the CoG shift due to pitch angle $\theta$, approximately $m_s g h_p \sin\theta$. For stability, the pitching moment from acceleration must not exceed the maximum restoring moment provided by the suspension’s pitch limit:
$$
m_s a_x h_p \leq m_s g h_p \sin(\theta_{\max})
$$
Thus, the longitudinal acceleration limit is:
$$
a_x \leq g \sin(\theta_{\max})
$$
For $\theta_{\max} = 6.6^\circ$, $a_x \leq 1.13 \text{ m/s}^2$.
Coupled Longitudinal and Lateral Acceleration Limit (Cornering): During cornering, lateral acceleration $a_y$ causes roll. The steady-state roll angle $\phi$ is approximated by:
$$
\phi = \frac{m_s h_r a_y}{k_\phi – m_s g h_p}
$$
where $k_\phi$ is the effective roll stiffness. The lateral acceleration is related to the longitudinal speed $v_x$ and path curvature $\rho$ (or turning radius $R$) by $a_y = v_x^2 \rho = v_x^2 / R$.
The most critical stability condition often occurs at the boundary of the CoG stable region. Considering the coupled effect, the condition for the CoG to remain stable can be expressed by combining the geometric constraint $y \geq -(B/L)x$ with the dynamic relationships for $x$ (due to pitch from $a_x$) and $y$ (due to roll from $a_y$). A conservative and practical limit for the maximum safe longitudinal speed during cornering can be derived as:
$$
v_x \leq \sqrt{ \frac{ \left(k_\phi – m_s g h_p\right) R}{m_s h_r} \left( h_p \theta_{\max} – \frac{B}{L} \right) }
$$
This formula provides the maximum speed for a given turn radius $R$ to ensure the dynamic CoG stays within the stability region. It is a crucial design constraint for the path-following controller of the fault-tolerant electric vehicle car. For example, with $k_\phi = 60000$ Nm/rad, $m_s=1260$ kg, $h_p=0.4$ m, $h_r=0.4$ m, $B/L=0.538$, $\theta_{\max}=0.115$ rad, and a turning radius of $R=50$ m, the theoretical maximum safe speed is approximately $32.4$ km/h.
2. Design of the Chassis-Coordinated Fault-Tolerant Controller for the Electric Vehicle Car
2.1 Overall Control Architecture
The proposed hierarchical control architecture for the corner-module electric vehicle car is designed to manage the complex, coupled dynamics of three-wheel driving. The architecture centers on the Active Suspension Controller (ASC), which is responsible for maintaining the lifted-wheel posture and managing the CoG position. It is supported by two optimized ancillary controllers that handle the consequences of the three-wheel configuration: a Traction Control System (TCS) to manage wheel slip and a Path Following Controller (PFC) to guide the vehicle. The interconnections are shown in the following description.
The Active Suspension Controller, based on Sliding Mode Control (SMC), calculates the required active force $F_{ai}$ for each of the four suspensions. It uses suspension travel sensors and reference positions (e.g., lift the faulty wheel to max extension, lower the diagonally opposite wheel) as inputs. The SMC design must account for strong coupling between suspensions.
The Traction Control System (TCS) is critical because lifting a wheel dynamically changes the vertical load on the other wheels, especially the wheel diagonally opposite the faulty one, which may experience reduced load and be prone to excessive slip during acceleration. A TCS based on Predictive Sliding Mode Control (PSMC) is designed. It uses wheel speeds, vehicle speed, and target slip ratios to calculate the desired drive torque $T_{di}$ for each functional in-wheel motor. The target slip ratio for the lightly loaded wheel is set conservatively (e.g., 0.2) to ensure grip and smooth power application.
The Path Following Controller (PFC) is based on Model Predictive Control (MPC). It takes the reference path (desired trajectory and curvature) and current vehicle states (yaw rate, sideslip, position error) as inputs. It calculates the required front and rear wheel steering angles ($\delta_f$, $\delta_r$) to minimize tracking error while respecting the vehicle’s dynamic constraints, notably the maximum lateral acceleration derived in Section 1.2. These angles are then distributed to the individual wheels of the electric vehicle car using Akermann steering geometry or similar allocation methods.
2.2 Active Suspension Controller (SMC) Design
The vertical dynamics of the electric vehicle car’s sprung mass, considering coupling between suspensions, can be modeled. For the design of the controller for suspension 1 (front-left), we define state variables related to its travel and velocity. The dynamic equation can be formulated as:
$$
\begin{aligned}
\dot{x}_{s1} &= x_{s2} \\
\dot{x}_{s2} &= S_{s1} + c_{u11} F_{a1}
\end{aligned}
$$
where $x_{s1}$ is the suspension travel, $x_{s2}$ is its velocity, $F_{a1}$ is the control input (active force), $c_{u11}$ is a constant parameter from the geometry and inertia properties, and $S_{s1}$ encapsulates the coupled effects from other suspensions’ forces and displacements, treated as a bounded disturbance ($|S_{s1}| \leq D_{s1}$).
Define the tracking error $e_{s1} = x_{s1} – x_{s1d}$, where $x_{s1d}$ is the desired suspension travel. Choose a sliding surface:
$s_{s1} = c e_{s1} + \dot{e}_{s1}$, where $c > 0$.
The derivative of the sliding surface is:
$$
\dot{s}_{s1} = c \dot{e}_{s1} + \ddot{e}_{s1} = c x_{s2} + S_{s1} + c_{u11} F_{a1} – \ddot{x}_{s1d}
$$
To achieve $\dot{s}_{s1} = -D_{s1} \text{sgn}(s_{s1}) – k s_{s1}$ (exponential reaching law), the control law is designed as:
$$
F_{a1} = \frac{1}{c_{u11}} \left( \ddot{x}_{s1d} – c x_{s2} – D_{s1} \text{sgn}(s_{s1}) – k s_{s1} \right)
$$
Using a Lyapunov function $V = \frac{1}{2} s_{s1}^2$, its derivative becomes $\dot{V} \leq -k V$, proving that the sliding surface $s_{s1}$ converges to zero exponentially, ensuring the suspension travel error converges. Controllers for the other three suspensions are designed similarly, enabling coordinated control to achieve the desired three-wheel stance for the electric vehicle car.
2.3 Traction Controller (PSMC) Design
A simplified quarter-car model is used for TCS design for each driven wheel of the electric vehicle car:
$$
\begin{aligned}
m_q \dot{v}_x &= F_x \\
J_\omega \dot{\omega} &= T_d – R_e F_x \\
\lambda &= \frac{\omega R_e – v_x}{\max(v_x, \omega R_e)} \quad \text{(Slip ratio definition)} \\
F_x &= F_z \mu(\lambda)
\end{aligned}
$$
where $m_q$ is a quarter vehicle mass, $v_x$ is the longitudinal velocity at the wheel, $\omega$ is the wheel angular speed, $J_\omega$ is the wheel inertia, $T_d$ is the drive torque, $R_e$ is the effective tire radius, $F_z$ is the vertical load, and $\mu(\lambda)$ is the tire-road friction coefficient as a function of slip ratio $\lambda$.
The derivative of the slip ratio can be expressed as:
$$
\dot{\lambda} = \frac{(1+\lambda)}{v_x} \left[ -\frac{F_z \mu(\lambda)}{m_q} \right] – \frac{R_e^2}{J_\omega v_x} F_z \mu(\lambda) + \frac{R_e}{J_\omega v_x} T_d
$$
Define the slip ratio error: $e_\lambda = \lambda – \lambda_d$, where $\lambda_d$ is the desired slip ratio (e.g., 0.2). Choose a sliding surface: $s_\lambda = c_\lambda e_\lambda + \dot{e}_\lambda$.
In Predictive Sliding Mode Control (PSMC), the control objective is to make the predicted value of the sliding surface $s_\lambda(t+T_\lambda)$ zero after a prediction horizon $T_\lambda$. This leads to the condition $s_\lambda(t) + T_\lambda \dot{s}_\lambda(t) = 0$.
Substituting the expression for $\dot{s}_\lambda$ and solving for the control input $T_d$ yields the PSMC law:
$$
T_d = \frac{J_\omega v_x}{R_e} \left[ \frac{\dot{\lambda}_d – c_\lambda e_\lambda – \frac{s_\lambda(t)}{T_\lambda}}{1+\lambda} + \frac{F_z \mu(\lambda)}{m_q v_x} + \frac{R_e^2}{J_\omega v_x} F_z \mu(\lambda) \right]
$$
This controller anticipates and counteracts slip, providing smooth and effective traction control for the electric vehicle car operating with three driven wheels.
2.4 Path Following Controller (MPC) Design
The lateral dynamics of the electric vehicle car for path tracking are described by a linear bicycle model augmented with path error states. The state-space model is:
$$
\begin{aligned}
\dot{\mathbf{x}}_c &= \mathbf{A}_c \mathbf{x}_c + \mathbf{B}_c \mathbf{u}_c + \mathbf{\omega}_c \\
\mathbf{y}_c &= \mathbf{C}_c \mathbf{x}_c
\end{aligned}
$$
where the state vector $\mathbf{x}_c = [v_y, \gamma, e_y, e_\psi]^T$ includes lateral velocity $v_y$, yaw rate $\gamma$, lateral path error $e_y$, and heading error $e_\psi$. The control input vector $\mathbf{u}_c = [\delta_f, \delta_r]^T$ consists of front and rear steering angles. The system matrices $\mathbf{A}_c$, $\mathbf{B}_c$, and $\mathbf{C}_c$ contain vehicle parameters (mass $m$, yaw inertia $I_z$, cornering stiffnesses $C_f$, $C_r$, wheelbase $L$, distances $a$, $b$). The disturbance $\mathbf{\omega}_c$ contains the path curvature $\rho$.
The MPC controller solves, at each time step, a finite-horizon optimal control problem:
$$
\min_{\mathbf{u}} \sum_{k=0}^{N-1} \left( \| \mathbf{y}_c(k) – \mathbf{y}_{ref}(k) \|^2_{\mathbf{Q}} + \| \mathbf{u}_c(k) \|^2_{\mathbf{R}} \right) + \| \mathbf{y}_c(N) – \mathbf{y}_{ref}(N) \|^2_{\mathbf{P}}
$$
subject to:
$$
\begin{aligned}
\mathbf{x}_c(k+1) &= \mathbf{A}_d \mathbf{x}_c(k) + \mathbf{B}_d \mathbf{u}_c(k) + \mathbf{\omega}_d(k) \quad \text{(Discretized dynamics)} \\
\mathbf{u}_{\min} &\leq \mathbf{u}_c(k) \leq \mathbf{u}_{\max} \quad \text{(Actuator limits)} \\
|a_y(k)| &= |v_x \gamma(k) + \dot{v}_y(k)| \leq a_{y,\max} \quad \text{(Lateral acceleration constraint)}
\end{aligned}
$$
Here, $\mathbf{Q}$, $\mathbf{R}$, $\mathbf{P}$ are weighting matrices. The critical lateral acceleration constraint $a_{y,\max}$ is calculated online based on the current speed $v_x$ and the stability limit derived in Section 1.2, ensuring the three-wheeled electric vehicle car remains stable during cornering. The MPC solves this quadratic programming problem to generate the optimal steering angles.
3. Simulation Validation and Analysis for the Electric Vehicle Car
The proposed fault-tolerant control strategy is validated using a high-fidelity co-simulation platform. The vehicle model is built in CarSim, which provides accurate vehicle dynamics, tire models (Magic Formula), and a detailed suspension model. The proposed chassis-coordinated controller is implemented in MATLAB/Simulink. The faulty scenario is a severely damaged (effectively flat) tire on the Rear-Right wheel (Wheel 4), reducing its rolling radius. The test maneuver is a lane-change (double lane change) at moderate speed, demanding both longitudinal and lateral control.
3.1 Simulation Results under Stable Control
The electric vehicle car starts from rest. The desired suspension travels are set to lift Wheel 4 and lower Wheel 1 to shift the CoG into the stable region: $z_{sd1}=140$ mm, $z_{sd2}=z_{sd3}=-140$ mm, $z_{sd4}=0$ mm. Drive torques are applied to Wheels 1, 2, and 3 to accelerate. Key results are summarized below.
Vehicle States and Stability: The active suspension controller successfully achieves the target suspension travels within about 1 second. Consequently, the vertical load on Wheel 4 drops to zero, confirming it is lifted, and the load on Wheel 1 also approaches zero initially as it is lowered. The vehicle attitude stabilizes with a pitch angle of about $5.4^\circ$ and a roll angle of $-8.8^\circ$, corresponding to a CoG position within the calculated stable trapezoidal region. The maximum longitudinal speed reached is $29.8$ km/h. During the lane-change maneuvers, lateral acceleration peaks are observed, but the vehicle remains stable. The CoG trajectory throughout the simulation stays strictly within the predefined stable region.
Traction and Path Following Performance: The PSMC-based TCS effectively manages wheel slip. The slip ratio of the critical Wheel 1 (diagonally opposite the faulty wheel) is maintained between $-0.22$ and $0.30$, preventing both excessive drive slip and lock-up. The MPC-based path-following controller demonstrates good performance. The maximum lateral tracking error during the demanding lane-change is $68.6$ mm, which is excellent for an electric vehicle car operating in a degraded three-wheel mode.
| Performance Metric | Simulation Result | Theoretical Limit / Target | Assessment |
|---|---|---|---|
| Max Achieved Speed | 29.8 km/h | 32.4 km/h (Theoretical Safe Speed for R=50m) | Close to limit, stable. |
| Max Lateral Tracking Error | 68.6 mm | Minimized by MPC | Very good performance. |
| Wheel 1 Max Slip Ratio | 0.30 | Target ~0.2 | Controlled within safe bound. |
| CoG Position | Remains inside stable region | Must remain inside stable region | Stability condition satisfied. |
| Faulty Wheel Load | ~0 N after 1.2s | 0 N (Lifted) | Successful fault isolation. |
3.2 Simulation Results Demonstrating Instability
To validate the theoretical speed limit, a second simulation is run where the drive torques are increased to force a higher speed during the same lane-change maneuver. The longitudinal speed reaches $36.9$ km/h, exceeding the theoretical safe speed of $32.4$ km/h.
Instability Onset: As the electric vehicle car enters the first turn at this higher speed, the increased lateral acceleration causes a larger dynamic load transfer. The CoG is pushed beyond the boundary of the stable region. This is immediately followed by a rapid change in vehicle attitude: the roll angle increases sharply, and the pitch angle decreases. The vertical load on Wheel 4 becomes non-zero, indicating it has contacted the ground again, while Wheel 1 loses contact. The vehicle transitions from a stable three-wheel configuration (on wheels 1,2,3) to an unstable, alternating contact state. The CoG trajectory is observed to leave the stable region, confirming the loss of the three-wheel stability condition.
This result critically validates the accuracy of the theoretical stability boundary analysis. The instability occurred at a speed ($36.9$ km/h) only $4.5$ km/h above the predicted safe limit ($32.4$ km/h), demonstrating that the derived model provides a reliable and necessary constraint for the safe operation of the three-wheeled electric vehicle car.
4. Conclusion and Future Work for the Fault-Tolerant Electric Vehicle Car
This paper has presented a comprehensive solution for a critical fault scenario in advanced corner-module electric vehicle cars: single wheel steering/drive system lock or tire damage. The proposed chassis-coordinated fault-tolerant control strategy enables the vehicle to continue its journey in a three-wheel driving mode, significantly enhancing its survivability and safety. The key contributions are:
- Stability Theory for Three-Wheel Driving: We derived the fundamental stability condition and the precise CoG stable region for a three-wheeled electric vehicle car, incorporating practical suspension travel limits. This analysis is foundational for any control strategy involving reduced wheel contact.
- Dynamic Safety Boundaries: We developed formulas for the maximum allowable longitudinal speed and lateral acceleration, considering their coupled effect on the CoG position. This provides a vital real-time constraint for the vehicle’s motion controller, ensuring operations remain within the stable domain.
- Integrated Hierarchical Controller Design: We designed a synergistic control architecture featuring a robust SMC-based active suspension controller as the core, supported by a PSMC-based traction controller and an MPC-based path follower. This coordination is essential for managing the interrelated challenges of posture control, wheel slip, and trajectory tracking in the degraded driving mode.
- Validation and Limit Verification: Co-simulation results confirmed the controller’s effectiveness, allowing stable path following at speeds close to the theoretical limit. Furthermore, simulations successfully induced instability by exceeding this limit, validating the accuracy of the derived stability model.
This research provides a solid theoretical foundation and a practical control framework for handling severe single-wheel faults in next-generation electric vehicle cars with corner module architectures. It transforms a scenario that would typically strand a vehicle into one where a controlled, safe retreat to a service location is possible.
Future work will focus on several important extensions:
- Experimental Validation: Implementing and testing the control strategy on a real corner-module electric vehicle car prototype is the crucial next step to assess practical performance, sensor noise effects, and actuator response delays.
- Extended Terrain and Conditions: The control strategy should be evaluated on uneven roads, low-friction surfaces, and during complex maneuvers to further robustify the algorithms.
- Actuator Dynamics and Fault Diagnosis: Incorporating detailed models of actuator (suspension actuator, in-wheel motor) dynamics and bandwidth limits into the controller design. Furthermore, integrating a real-time fault detection and isolation (FDI) module to automatically identify the faulty wheel and trigger the appropriate control mode.
- Multi-Fault Scenarios: Investigating the feasibility and strategies for handling concurrent faults in more than one corner module, which would present even greater challenges for the electric vehicle car’s fault-tolerant capabilities.