控制工具PID

最新推荐文章于 2025-08-22 13:59:23 发布

原创最新推荐文章于 2025-08-22 13:59:23 发布 · 1.4k 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#c++ #机器人工程 #机器人编程

本文介绍了阻尼振荡器的工作原理，并通过示例展示了如何使用PID控制器进行控制。代码示例中定义了一个自定义控制器类，包含比例(P)和微分(D)增益，用于在二阶系统上实现控制。通过设置控制器参数并整合到系统中，最终实现了对阻尼振荡器状态的调整。

前一篇：

pressbooks.online.ucf.edu/phy2048tjb/chapter/15-5-damped-oscillations/

byjus.com/jee/damped-oscillation/

阻尼振荡器原理上面两个链接，程序下面一个链接。

控制工具基础案例

如何控制呢？PID？

看的如何很懵，正常，后续会把这些点串起来。

/**********************************************************************************************************************
This file is part of the Control Toolbox (https://github.com/ethz-adrl/control-toolbox), copyright by ETH Zurich.
Licensed under the BSD-2 license (see LICENSE file in main directory)
 **********************************************************************************************************************/
#pragma once
class CustomController : public ct::core::Controller<2, 1>
{
public:
    static const size_t state_dim = 2;    // two states
    static const size_t control_dim = 1;  // one control action
    CustomController(const ct::core::ControlVector<control_dim>& uff,  // feedforward control
        const double& kp,                                              // P gain
        const double& kd                                               // D gain
        )
        : uff_(uff), kp_(kp), kd_(kd)
    {
    }
    ~CustomController() {}
    CustomController(const CustomController& other) : uff_(other.uff_), kp_(other.kp_), kd_(other.kd_) {}
    CustomController* clone() const override
    {
        return new CustomController(*this);  // calls copy constructor
    }
    void computeControl(const ct::core::StateVector<state_dim>& state,
        const double& t,
        ct::core::ControlVector<control_dim>& controlAction) override
    {
        controlAction = uff_;                                 // apply feedforward control
        controlAction(0) -= kp_ * state(0) + kd_ * state(1);  // add feedback control
    }
private:
    ct::core::ControlVector<control_dim> uff_;  
    double kp_;                                 
    double kd_;                                 
};

/**********************************************************************************************************************
This file is part of the Control Toolbox (https://github.com/ethz-adrl/control-toolbox), copyright by ETH Zurich.
Licensed under the BSD-2 license (see LICENSE file in main directory)
**********************************************************************************************************************/
#include <ct/core/core.h>
#include <ct/core/examples/CustomController.h>
int main(int argc, char** argv)
{
    // a damped oscillator has two states, position and velocity
    const size_t state_dim = ct::core::SecondOrderSystem::STATE_DIM;      // = 2
    const size_t control_dim = ct::core::SecondOrderSystem::CONTROL_DIM;  // = 1
    // create a state
    ct::core::StateVector<state_dim> x;
    // we initialize it at a point with unit deflection and zero velocity
    x(0) = 1.0;
    x(1) = 0.0;
    // create our oscillator
    double w_n = 50;
    std::shared_ptr<ct::core::SecondOrderSystem> oscillator(new ct::core::SecondOrderSystem(w_n));
    // create our controller
    double kp = 10;
    double kd = 1;
    ct::core::ControlVector<control_dim> uff;
    uff << 2.0;
    std::shared_ptr<CustomController> controller(new CustomController(uff, kp, kd));
    // assign our controller
    oscillator->setController(controller);
    // create an integrator
    ct::core::Integrator<state_dim> integrator(oscillator, ct::core::IntegrationType::RK4);
    // simulate 1000 steps
    double dt = 0.001;
    ct::core::Time t0 = 0.0;
    size_t nSteps = 1000;
    integrator.integrate_n_steps(x, t0, nSteps, dt);
    // print the new state
    std::cout << "state after integration: " << x.transpose() << std::endl;
    return 0;
}