Recursive Newton-Euler algorithm

MatLab implementation of recursive Newton-Euler algorithm for robot dynamics. This script contains the implementation alongside with three examples how to use these functions.

Implementation details

In order to use the functions correctly, one needs to define a data structure that describes the physical and geometrical parameters of the robot.

robot struct contatins following parameters:

• njoints - number of joints
• type - array of 'R' or 'P' values describing the type of a joint, starting from the first one
• mass - array of masses of each link, starting from the first one
• length - array of lengthes of each link, starting from the first one
• g - three-dimensional array describing the axis where the gravity acts
• R - cell array of rotation matrices from frame {i-1} to {i} starting from $R^0_1$
• inertia - cell array of 3x3 inertia matrices of each link, starting from the first one.

RNE function takes as arguments following values:

• q - array of joint values
• qdot - array of joint velocities
• q2dot - array of joint accelerations
• axis - 3x1 unit vector that describes alongside what axis the position vector from origin of {i} to origin of {i-1}

First example

A planar 2R robot is given with the following parameters:

• q = [90, 45]
• qdot= [-0.8, 0.35]
• q2dot = [-0.4, 0.1]

Second example

A planar RP robot with given parameters:

• q = [20, 0.2]
• qdot = [0.08, 0.03]
• q2dot = [0.1, 0.01]

Third example

3R robot with given parameters:

• q = [20, 40, 10]
• qdot = [0.2, 0.15, -0.2]
• q2dot = [0.1, 0.085, 0]