mirror of
https://github.com/AtsushiSakai/PythonRobotics.git
synced 2026-04-22 03:00:22 -04:00
107 lines
3.3 KiB
ReStructuredText
107 lines
3.3 KiB
ReStructuredText
|
|
Quintic polynomials planning
|
|
----------------------------
|
|
|
|
Motion planning with quintic polynomials.
|
|
|
|
.. image:: https://github.com/AtsushiSakai/PythonRoboticsGifs/raw/master/PathPlanning/QuinticPolynomialsPlanner/animation.gif
|
|
|
|
It can calculate 2D path, velocity, and acceleration profile based on
|
|
quintic polynomials.
|
|
|
|
|
|
|
|
Quintic polynomials for one dimensional robot motion
|
|
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
|
|
We assume a one-dimensional robot motion :math:`x(t)` at time :math:`t` is
|
|
formulated as a quintic polynomials based on time as follows:
|
|
|
|
.. math:: x(t) = a_0+a_1t+a_2t^2+a_3t^3+a_4t^4+a_5t^5
|
|
:label: eq1
|
|
|
|
:math:`a_0, a_1. a_2, a_3, a_4, a_5` are parameters of the quintic polynomial.
|
|
|
|
It is assumed that terminal states (start and end) are known as boundary conditions.
|
|
|
|
Start position, velocity, and acceleration are :math:`x_s, v_s, a_s` respectively.
|
|
|
|
End position, velocity, and acceleration are :math:`x_e, v_e, a_e` respectively.
|
|
|
|
So, when time is 0.
|
|
|
|
.. math:: x(0) = a_0 = x_s
|
|
:label: eq2
|
|
|
|
Then, differentiating the equation :eq:`eq1` with t,
|
|
|
|
.. math:: x'(t) = a_1+2a_2t+3a_3t^2+4a_4t^3+5a_5t^4
|
|
:label: eq3
|
|
|
|
So, when time is 0,
|
|
|
|
.. math:: x'(0) = a_1 = v_s
|
|
:label: eq4
|
|
|
|
Then, differentiating the equation :eq:`eq3` with t again,
|
|
|
|
.. math:: x''(t) = 2a_2+6a_3t+12a_4t^2
|
|
:label: eq5
|
|
|
|
So, when time is 0,
|
|
|
|
.. math:: x''(0) = 2a_2 = a_s
|
|
:label: eq6
|
|
|
|
so, we can calculate :math:`a_0, a_1, a_2` with eq. :eq:`eq2`, :eq:`eq4`, :eq:`eq6` and boundary conditions.
|
|
|
|
:math:`a_3, a_4, a_5` are still unknown in eq :eq:`eq1`.
|
|
|
|
We assume that the end time for a maneuver is :math:`T`, we can get these equations from eq :eq:`eq1`, :eq:`eq3`, :eq:`eq5`:
|
|
|
|
.. math:: x(T)=a_0+a_1T+a_2T^2+a_3T^3+a_4T^4+a_5T^5=x_e
|
|
:label: eq7
|
|
|
|
.. math:: x'(T)=a_1+2a_2T+3a_3T^2+4a_4T^3+5a_5T^4=v_e
|
|
:label: eq8
|
|
|
|
.. math:: x''(T)=2a_2+6a_3T+12a_4T^2+20a_5T^3=a_e
|
|
:label: eq9
|
|
|
|
From eq :eq:`eq7`, :eq:`eq8`, :eq:`eq9`, we can calculate :math:`a_3, a_4, a_5` to solve the linear equations: :math:`Ax=b`
|
|
|
|
.. math:: \begin{bmatrix} T^3 & T^4 & T^5 \\ 3T^2 & 4T^3 & 5T^4 \\ 6T & 12T^2 & 20T^3 \end{bmatrix}\begin{bmatrix} a_3\\ a_4\\ a_5\end{bmatrix}=\begin{bmatrix} x_e-x_s-v_sT-0.5a_sT^2\\ v_e-v_s-a_sT\\ a_e-a_s\end{bmatrix}
|
|
|
|
We can get all unknown parameters now.
|
|
|
|
Quintic polynomials for two dimensional robot motion (x-y)
|
|
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
|
|
If you use two quintic polynomials along x axis and y axis, you can plan for two dimensional robot motion in x-y plane.
|
|
|
|
.. math:: x(t) = a_0+a_1t+a_2t^2+a_3t^3+a_4t^4+a_5t^5
|
|
:label: eq10
|
|
|
|
.. math:: y(t) = b_0+b_1t+b_2t^2+b_3t^3+b_4t^4+b_5t^5
|
|
:label: eq11
|
|
|
|
It is assumed that terminal states (start and end) are known as boundary conditions.
|
|
|
|
Start position, orientation, velocity, and acceleration are :math:`x_s, y_s, \theta_s, v_s, a_s` respectively.
|
|
|
|
End position, orientation, velocity, and acceleration are :math:`x_e, y_e. \theta_e, v_e, a_e` respectively.
|
|
|
|
Each velocity and acceleration boundary condition can be calculated with each orientation.
|
|
|
|
:math:`v_{xs}=v_scos(\theta_s), v_{ys}=v_ssin(\theta_s)`
|
|
|
|
:math:`v_{xe}=v_ecos(\theta_e), v_{ye}=v_esin(\theta_e)`
|
|
|
|
References:
|
|
~~~~~~~~~~~
|
|
|
|
- `Local Path Planning And Motion Control For Agv In
|
|
Positioning <http://ieeexplore.ieee.org/document/637936/>`__
|
|
|
|
|