Actuator sizing of four wheel differential robot on simScape multibody

Dear Matlab community,

I have a problem on torque measurment in an imported urdf model for a differential drive model as shown below.

The picture below illustrate the 3D model of the robot to be simulated.

I have tried for eternity to ajdust the spatial contact normal force parameter, the latest equation used to calculate the stiffness and damping

the above photo illustrate the configuration of spatial contact between the wheel and the ground.

The wheel is given an acceleration signal equal to 6.25 m/sec^2 and I want to measure the torque of the joint, but when I measure the torque the output is non-sense and noisy.

Like this, this is the torque measurment for the joint in N.m it's very noise and non-sense and I can't figure out how to solve it or how to find the problem.

I am using matlab R2023a

So if there any advice or solution on how to model it correctly or how to solve it, I will apperciate it very much

[Chuguang Pan]

I think the problem is the value of stiffness is too large. The large stiffness causes the simulation solver select very small time step to integrate the overall system.

[Jan Janse van Rensburg]

There is a few things that can play a role here, what solver are you using? What is the wheel inertias and how are they calculated. What are you using for the contact is it an extruded or revolved solid or a CAD-file of the wheel or a primitive cylinder? Your friction factors also seem quite small, so with a large torque you may be oscillating between stick and slip. Firstly make sure you are using realistic values for the mass and inertias of the wheels and body, secondly switch to a variable step solver (ode23t or similar). It. May also be worthwhile to add a marker on the wheel (add a massless block or offset cylinder so you can see the rotation of the wheels - they may be spinning quite fast). That noise seems like either a solver issue or a badly parameterized model.