Abstract:This paper proposes a data-driven model predictive control for multirotor collision avoidance considering uncertainty and an unknown model from a payload. To address this challenge, sparse identification of nonlinear dynamics (SINDy) is used to obtain the governing equation of the multirotor system. The SINDy can discover the equations of target systems with low data, assuming that few functions have the dominant characteristic of the system. Model predictive control (MPC) is utilized to obtain accurate trajectory tracking performance by considering state and control input constraints. To avoid a collision during operation, MPC optimization problem is again formulated using inequality constraints about an obstacle. In simulation, SINDy can discover a governing equation of multirotor system including mass parameter uncertainty and aerodynamic effects. In addition, the simulation results show that the proposed method has the capability to avoid an obstacle and track the desired trajectory accurately.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the challenge of collision avoidance and trajectory tracking for multi - rotor unmanned aerial vehicles (UAVs) in the presence of uncertainties and unknown models (such as the influence from the load). Specifically, the author proposes a data - driven model predictive control (MPC) method based on Sparse Identification of Nonlinear Dynamics (SINDy) to improve the trajectory - tracking performance of multi - rotor UAVs in autonomous flight and ensure that they can effectively avoid obstacles.
### Main problems
1. **Uncertainties and unknown models**: When a multi - rotor UAV carries a load, its mass parameters and aerodynamic effects will change. These changes will affect the dynamic characteristics of the system, making it difficult for traditional control methods to accurately track the desired trajectory.
2. **Collision avoidance**: When performing tasks in a complex environment, the UAV needs to be able to sense and avoid obstacles in real - time, which places higher requirements on the control algorithm.
### Solutions
To solve the above problems, the method proposed in this paper includes the following key steps:
- **Using SINDy to obtain system equations**: Identify the nonlinear differential equations describing the behavior of the multi - rotor system from a small amount of data through the SINDy method. SINDy assumes that the main characteristics of the system are dominated by a few functions, so the governing equations of the system can be obtained with less data.
\[
\dot{x} = f(x, u)
\]
where \(x\in\mathbb{R}^m\) is the state vector, \(u\in\mathbb{R}^n\) is the control input vector, and \(f\) represents the nonlinear system function.
- **Combining MPC for optimal control**: Use the model obtained by SINDy to design an MPC controller to calculate the optimal control input. MPC minimizes the cost function by solving an optimization problem that includes state and control input constraints:
\[
\min_{u_1,\cdots,u_N}\sum_{k = 1}^{N}\left((r_i - y_i)^TQ(r_i - y_i)+u_i^TRu_i\right)
\]
The constraints include:
\[
\begin{aligned}
&\dot{x}=\hat{f}(x, u),\\
&u_{\text{min}}\leq u\leq u_{\text{max}},\\
&x_1 = x_{\text{init}},\\
&(x_{\text{ob}}-x)^2+(y_{\text{ob}}-y)^2+(z_{\text{ob}}-z)^2\geq D_{\text{min}},
\end{aligned}
\]
where \(r\) is the reference trajectory, \(\hat{f}\) is the data - driven multi - rotor model obtained from SINDy, \(u_{\text{min}}\) and \(u_{\text{max}}\) are the lower and upper limits of the control input, \(x_{\text{init}}\) is the initial state, \(Q\) and \(R\) are the weight matrices of the error vector and the control input vector respectively, \(N\) is the moving horizon step size, \(x_{\text{ob}}, y_{\text{ob}}, z_{\text{ob}}\) are the obstacle positions, and \(D_{\text{min}}\) is the minimum distance to ensure safety.
### Summary
The SINDy - MPC method proposed in this paper can not only handle the parameter uncertainties caused by the load and the influence of the unknown aerodynamic model, but also effectively achieve collision avoidance, thereby improving the autonomous flight ability of multi - rotor UAVs in complex environments.