Abstract:Vehicle speed, road roughness grade and sprung mass are the three main factors to influence suspension control and state estimation. Aiming at the problem that fixed state observer cannot guarantee the estimation accuracy of suspension with driving scenario changes, a suspension state observer based on interactive multiple model adaptive Kalman filter (IMMAKF) is established. Firstly, an adaptive control suspension is proposed based on LQR algorithm and multi-objective optimization algorithm, which can automatically adjust the controller parameters according to the vehicle speed, road roughness grade and sprung acceleration parameters, so as to keep the optimal control effect of the suspension. Secondly, the theoretical model of IMMAKF is derived, and two kinds of IMMAKF suspension state observers and controllers are established. Finally, a simulation condition with the vehicle speed, road roughness grade and sprung mass changing simultaneously is established. The simulation results shows that: compared with ordinary IMMKF, AKF and KF observers, the estimation accuracy of IMMAKF5 is improved. Except for state observation, IMMAKF can be used to identify the road roughness grade and estimate the suspension sprung mass.
What problem does this paper attempt to address?
This paper aims to solve the problem of low state estimation accuracy of the adaptive suspension system under different driving conditions. Specifically, traditional fixed - state observers cannot ensure the state estimation accuracy of the suspension system when the driving scenario changes. To solve this problem, the paper proposes a suspension state observer based on the Interacting Multiple - Model Adaptive Kalman Filter (IMMAKF).
### Main problems
1. **Limitations of fixed - state observers**:
- Traditional fixed - state observers cannot maintain high - accuracy state estimation when driving conditions change.
- This leads to a decline in the suspension control effect, especially under different vehicle speeds, road roughness, and sprung - mass conditions.
2. **Requirements for adaptive suspension control**:
- As the vehicle speed, road roughness, and sprung - mass change, the vibration performance of the suspension system will be affected.
- Therefore, a method that can automatically adjust the controller parameters according to these changes is required to maintain the optimal suspension control effect.
### Solutions
The paper proposes the following solutions:
1. **Adaptive control suspension based on LQR algorithm and multi - objective optimization algorithm**:
- An adaptive control suspension based on the Linear Quadratic Regulator (LQR) algorithm and multi - objective optimization algorithm is proposed.
- This suspension can automatically adjust the controller parameters according to vehicle speed, road roughness, and sprung - acceleration parameters to maintain the optimal control effect of the suspension.
2. **Derivation of the IMMAKF theoretical model**:
- The theoretical model of IMMAKF is derived, and two types of IMMAKF suspension state observers and controllers are established.
- IMMAKF improves the state estimation accuracy by combining the Adaptive Kalman Filter (AKF) and the Interacting Multiple - Model Kalman Filter (IMMKF).
3. **Simulation verification**:
- A simulation condition in which the vehicle speed, road roughness, and sprung - mass are simultaneously changed is established.
- The simulation results show that, compared with ordinary IMMKF, AKF, and KF observers, the estimation accuracy of IMMAKF is improved.
### Main contributions
1. **Proposing the IMMAKF suspension state estimation theory**:
- The IMMAKF observer can improve the state estimation accuracy when the vehicle speed, road grade, and sprung - mass change.
2. **Proposing a new sprung - mass estimation theory**:
- The sprung - mass is estimated through the interaction probability of the IMMAKF model without calculating it as one of the state parameters.
3. **Proposing a new adaptive suspension control system**:
- This control system can improve ride comfort under different working conditions.
### Conclusion
By proposing the IMMAKF observer, the paper solves the problem of low state estimation accuracy of traditional fixed - state observers when driving conditions change, providing new methods and ideas for the design of adaptive suspension systems.