I am trying to set up a 4-phase "Hybrid Stepper Motor". When I use a two-phase motor, I am able to connect it according to the documentation. However, when I try to use the four-phase variant, I see two additional ports "CA" and "CB" appear. However, I cannot find references on these ports or how to connect them in the "Stepper Motor" documentation. What is the physical interpretation of these ports?I am trying to set up a 4-phase "Hybrid Stepper Motor". When I use a two-phase motor, I am able to connect it according to the documentation. However, when I try to use the four-phase variant, I see two additional ports "CA" and "CB" appear. However, I cannot find references on these ports or how to connect them in the "Stepper Motor" documentation. What is the physical interpretation of these ports? I am trying to set up a 4-phase "Hybrid Stepper Motor". When I use a two-phase motor, I am able to connect it according to the documentation. However, when I try to use the four-phase variant, I see two additional ports "CA" and "CB" appear. However, I cannot find references on these ports or how to connect them in the "Stepper Motor" documentation. What is the physical interpretation of these ports? phasewinding, specializedpowersystems, simscapeelectrical, steppermotor MATLAB Answers — New Questions

​

Hello, I have to read a Audio File (.wav) into a vector.
I know how to plot it:
info = audioinfo(‘sound.wav’);
[y, Fs] = audioread(‘sound.wav’);

t = 0:1/Fs:info.Duration;
t = t(1:end-1);

plot(t,y);
xlabel(‘Time’);
ylabel(‘Audio Signal’);
But this Sript does not read the Audio into a Vector.
I also tried to read every single Sample, but this is to complex and I get error after error. Is there a simple way to do this?Hello, I have to read a Audio File (.wav) into a vector.
I know how to plot it:
info = audioinfo(‘sound.wav’);
[y, Fs] = audioread(‘sound.wav’);

t = 0:1/Fs:info.Duration;
t = t(1:end-1);

plot(t,y);
xlabel(‘Time’);
ylabel(‘Audio Signal’);
But this Sript does not read the Audio into a Vector.
I also tried to read every single Sample, but this is to complex and I get error after error. Is there a simple way to do this? Hello, I have to read a Audio File (.wav) into a vector.
I know how to plot it:
info = audioinfo(‘sound.wav’);
[y, Fs] = audioread(‘sound.wav’);

t = 0:1/Fs:info.Duration;
t = t(1:end-1);

plot(t,y);
xlabel(‘Time’);
ylabel(‘Audio Signal’);
But this Sript does not read the Audio into a Vector.
I also tried to read every single Sample, but this is to complex and I get error after error. Is there a simple way to do this? vector, audio, sound MATLAB Answers — New Questions

​

Post Content Post Content dfig npc converter MATLAB Answers — New Questions

​

I wrote a MWE to compare the perforance of interp1 (MATLAB built-in) with my custom interp1_scal (for linear interpolation), using a simple dynamic programming problem solved with value function iteration
$$ V(k) = max_{k’} log(k^alpha-k’)+beta*V(k’) $$
Question 1: I am a bit surprised that the custom interpolation routine I wrote is much faster than the built-in interp1 for linear interpolation (on my laptop with R2024b, it is 4 times faster, here (RUN button) the speed advantage is smaller but still significant). To find the location of the query point xi on the x grid, I am using binary search (see function locate), directly translated from an older Fortran/C routine.
Question 2: Am I using interp1 in the wrong way? would using griddedInterp or other Matlab routines significantly improve run times of this MWE?
Any help, suggestions or comments is greatly appreciated!
P.S. I am glad to provide additional info on the dynamic programming problem that this simple MWE is meant to caputure in few lines of code, if someone asks for.
Minimum Working Example here:
clear,clc,close all

%% Parameters
maxit = 50;
tol = 1e-5;
alpha = 0.33;
beta = 0.96;
k_ss = (alpha * beta)^(1 / (1 – alpha));
n_k = 300;
k_min = 0.5 * k_ss;
k_max = 1.5 * k_ss;
k_grid = linspace(k_min, k_max, n_k)’;
% Precompute output y = k^alpha
y_grid = k_grid .^ alpha;

%% Method 1: value function iteration with Matlab built-in interpolation
tic
% Initialization
V0 = zeros(n_k, 1);
V1 = zeros(n_k, 1);

% Value Function Iteration
for iter=1:maxit
for i = 1:n_k
y_val = y_grid(i);
k_max_now = min(y_val, k_max); % upper bound for Brent minimization
% Objective function (negative value for minimization)
obj = @(kp) -(log(y_val-kp) + beta*interp1(k_grid,V0,kp,’linear’));
% Optimal policy
kp_opt = fminbnd(obj, k_min, k_max_now);
% Updated value function
V1(i) = -obj(kp_opt);
end
% Convergence check
err = max(abs(V1 – V0));
V0 = V1;
end %end iter
time_vfi1=toc;

%% Method 2: value function iteration with *custom* interpolation
tic
% Initialization
V0 = zeros(n_k, 1);
V2 = zeros(n_k, 1);

% Value Function Iteration
for iter=1:maxit
for i_k = 1:n_k
y_val = y_grid(i_k);
k_max_now = min(y_val, k_max);
% Objective function (negative value for minimization)
obj = @(kp) -( log(y_val-kp) + beta*interp1_scal(k_grid,V0,kp) );
% Optimal policy
kp_opt = fminbnd(obj, k_min, k_max_now);
% Updated value function
V2(i_k) = -obj(kp_opt);
end
% Convergence check
err = max(abs(V2 – V0));
V0 = V2;
end %end iter
time_vfi2=toc;

%% Compare vfi1 and vfi2
err_V = max(abs(V1-V2));
disp(‘RESULTS OF COMPARISON’)
fprintf(‘||V1-V2|| = %f n’,err_V)
fprintf(‘Run time of vfi1 with Matlab interp = %f n’,time_vfi1)
fprintf(‘Run time of vfi2 with custom interp = %f n’,time_vfi2)
fprintf(‘time vfi1 / time_vfi2 = %f n’,time_vfi1/time_vfi2)

%% Plot value function
figure;
plot(k_grid, V1, ‘LineWidth’, 1.5);
hold on
plot(k_grid, V2, ‘LineWidth’, 1.5);
legend(‘V1′,’V2’)
xlabel(‘Current capital’);
grid on;

% Fast linear interpolation routine
% Usage:
% yi = interp1_scal(x,y,xi)
% where x and y are column vectors with n elements, xi is a scalar
% and yi is a scalar
% Input Arguments
% x – Sample points
% column vector
% Y – Sample data
% column vector
% xi – Query point
% scalar
function yi = interp1_scal(x,y,xi)

n = size(x,1);
j = locate(x,xi);
j = max(min(j,n-1),1);

slope = (y(j+1)-y(j))/(x(j+1)-x(j));
yi = y(j)+(xi-x(j))*slope;

end %end function interp1_scal

function jl = locate(xx,x)
%function jl = locate(xx,x)
%
% x is between xx(jl) and xx(jl+1)
%
% jl = 0 and jl = n means x is out of range
%
% xx is assumed to be monotone increasing

n = length(xx);
if x<xx(1)
jl = 0;
elseif x>xx(n)
jl = n;
else
jl = 1;
ju = n;
while (ju-jl>1)
jm = floor((ju+jl)/2);
if x>=xx(jm)
jl = jm;
else
ju=jm;
end
end
end

end %end function locateI wrote a MWE to compare the perforance of interp1 (MATLAB built-in) with my custom interp1_scal (for linear interpolation), using a simple dynamic programming problem solved with value function iteration
$$ V(k) = max_{k’} log(k^alpha-k’)+beta*V(k’) $$
Question 1: I am a bit surprised that the custom interpolation routine I wrote is much faster than the built-in interp1 for linear interpolation (on my laptop with R2024b, it is 4 times faster, here (RUN button) the speed advantage is smaller but still significant). To find the location of the query point xi on the x grid, I am using binary search (see function locate), directly translated from an older Fortran/C routine.
Question 2: Am I using interp1 in the wrong way? would using griddedInterp or other Matlab routines significantly improve run times of this MWE?
Any help, suggestions or comments is greatly appreciated!
P.S. I am glad to provide additional info on the dynamic programming problem that this simple MWE is meant to caputure in few lines of code, if someone asks for.
Minimum Working Example here:
clear,clc,close all

%% Parameters
maxit = 50;
tol = 1e-5;
alpha = 0.33;
beta = 0.96;
k_ss = (alpha * beta)^(1 / (1 – alpha));
n_k = 300;
k_min = 0.5 * k_ss;
k_max = 1.5 * k_ss;
k_grid = linspace(k_min, k_max, n_k)’;
% Precompute output y = k^alpha
y_grid = k_grid .^ alpha;

%% Method 1: value function iteration with Matlab built-in interpolation
tic
% Initialization
V0 = zeros(n_k, 1);
V1 = zeros(n_k, 1);

% Value Function Iteration
for iter=1:maxit
for i = 1:n_k
y_val = y_grid(i);
k_max_now = min(y_val, k_max); % upper bound for Brent minimization
% Objective function (negative value for minimization)
obj = @(kp) -(log(y_val-kp) + beta*interp1(k_grid,V0,kp,’linear’));
% Optimal policy
kp_opt = fminbnd(obj, k_min, k_max_now);
% Updated value function
V1(i) = -obj(kp_opt);
end
% Convergence check
err = max(abs(V1 – V0));
V0 = V1;
end %end iter
time_vfi1=toc;

%% Method 2: value function iteration with *custom* interpolation
tic
% Initialization
V0 = zeros(n_k, 1);
V2 = zeros(n_k, 1);

% Value Function Iteration
for iter=1:maxit
for i_k = 1:n_k
y_val = y_grid(i_k);
k_max_now = min(y_val, k_max);
% Objective function (negative value for minimization)
obj = @(kp) -( log(y_val-kp) + beta*interp1_scal(k_grid,V0,kp) );
% Optimal policy
kp_opt = fminbnd(obj, k_min, k_max_now);
% Updated value function
V2(i_k) = -obj(kp_opt);
end
% Convergence check
err = max(abs(V2 – V0));
V0 = V2;
end %end iter
time_vfi2=toc;

%% Compare vfi1 and vfi2
err_V = max(abs(V1-V2));
disp(‘RESULTS OF COMPARISON’)
fprintf(‘||V1-V2|| = %f n’,err_V)
fprintf(‘Run time of vfi1 with Matlab interp = %f n’,time_vfi1)
fprintf(‘Run time of vfi2 with custom interp = %f n’,time_vfi2)
fprintf(‘time vfi1 / time_vfi2 = %f n’,time_vfi1/time_vfi2)

%% Plot value function
figure;
plot(k_grid, V1, ‘LineWidth’, 1.5);
hold on
plot(k_grid, V2, ‘LineWidth’, 1.5);
legend(‘V1′,’V2’)
xlabel(‘Current capital’);
grid on;

% Fast linear interpolation routine
% Usage:
% yi = interp1_scal(x,y,xi)
% where x and y are column vectors with n elements, xi is a scalar
% and yi is a scalar
% Input Arguments
% x – Sample points
% column vector
% Y – Sample data
% column vector
% xi – Query point
% scalar
function yi = interp1_scal(x,y,xi)

n = size(x,1);
j = locate(x,xi);
j = max(min(j,n-1),1);

slope = (y(j+1)-y(j))/(x(j+1)-x(j));
yi = y(j)+(xi-x(j))*slope;

end %end function interp1_scal

function jl = locate(xx,x)
%function jl = locate(xx,x)
%
% x is between xx(jl) and xx(jl+1)
%
% jl = 0 and jl = n means x is out of range
%
% xx is assumed to be monotone increasing

n = length(xx);
if x<xx(1)
jl = 0;
elseif x>xx(n)
jl = n;
else
jl = 1;
ju = n;
while (ju-jl>1)
jm = floor((ju+jl)/2);
if x>=xx(jm)
jl = jm;
else
ju=jm;
end
end
end

end %end function locate I wrote a MWE to compare the perforance of interp1 (MATLAB built-in) with my custom interp1_scal (for linear interpolation), using a simple dynamic programming problem solved with value function iteration
$$ V(k) = max_{k’} log(k^alpha-k’)+beta*V(k’) $$
Question 1: I am a bit surprised that the custom interpolation routine I wrote is much faster than the built-in interp1 for linear interpolation (on my laptop with R2024b, it is 4 times faster, here (RUN button) the speed advantage is smaller but still significant). To find the location of the query point xi on the x grid, I am using binary search (see function locate), directly translated from an older Fortran/C routine.
Question 2: Am I using interp1 in the wrong way? would using griddedInterp or other Matlab routines significantly improve run times of this MWE?
Any help, suggestions or comments is greatly appreciated!
P.S. I am glad to provide additional info on the dynamic programming problem that this simple MWE is meant to caputure in few lines of code, if someone asks for.
Minimum Working Example here:
clear,clc,close all

%% Parameters
maxit = 50;
tol = 1e-5;
alpha = 0.33;
beta = 0.96;
k_ss = (alpha * beta)^(1 / (1 – alpha));
n_k = 300;
k_min = 0.5 * k_ss;
k_max = 1.5 * k_ss;
k_grid = linspace(k_min, k_max, n_k)’;
% Precompute output y = k^alpha
y_grid = k_grid .^ alpha;

%% Method 1: value function iteration with Matlab built-in interpolation
tic
% Initialization
V0 = zeros(n_k, 1);
V1 = zeros(n_k, 1);

% Value Function Iteration
for iter=1:maxit
for i = 1:n_k
y_val = y_grid(i);
k_max_now = min(y_val, k_max); % upper bound for Brent minimization
% Objective function (negative value for minimization)
obj = @(kp) -(log(y_val-kp) + beta*interp1(k_grid,V0,kp,’linear’));
% Optimal policy
kp_opt = fminbnd(obj, k_min, k_max_now);
% Updated value function
V1(i) = -obj(kp_opt);
end
% Convergence check
err = max(abs(V1 – V0));
V0 = V1;
end %end iter
time_vfi1=toc;

%% Method 2: value function iteration with *custom* interpolation
tic
% Initialization
V0 = zeros(n_k, 1);
V2 = zeros(n_k, 1);

% Value Function Iteration
for iter=1:maxit
for i_k = 1:n_k
y_val = y_grid(i_k);
k_max_now = min(y_val, k_max);
% Objective function (negative value for minimization)
obj = @(kp) -( log(y_val-kp) + beta*interp1_scal(k_grid,V0,kp) );
% Optimal policy
kp_opt = fminbnd(obj, k_min, k_max_now);
% Updated value function
V2(i_k) = -obj(kp_opt);
end
% Convergence check
err = max(abs(V2 – V0));
V0 = V2;
end %end iter
time_vfi2=toc;

%% Compare vfi1 and vfi2
err_V = max(abs(V1-V2));
disp(‘RESULTS OF COMPARISON’)
fprintf(‘||V1-V2|| = %f n’,err_V)
fprintf(‘Run time of vfi1 with Matlab interp = %f n’,time_vfi1)
fprintf(‘Run time of vfi2 with custom interp = %f n’,time_vfi2)
fprintf(‘time vfi1 / time_vfi2 = %f n’,time_vfi1/time_vfi2)

%% Plot value function
figure;
plot(k_grid, V1, ‘LineWidth’, 1.5);
hold on
plot(k_grid, V2, ‘LineWidth’, 1.5);
legend(‘V1′,’V2’)
xlabel(‘Current capital’);
grid on;

% Fast linear interpolation routine
% Usage:
% yi = interp1_scal(x,y,xi)
% where x and y are column vectors with n elements, xi is a scalar
% and yi is a scalar
% Input Arguments
% x – Sample points
% column vector
% Y – Sample data
% column vector
% xi – Query point
% scalar
function yi = interp1_scal(x,y,xi)

n = size(x,1);
j = locate(x,xi);
j = max(min(j,n-1),1);

slope = (y(j+1)-y(j))/(x(j+1)-x(j));
yi = y(j)+(xi-x(j))*slope;

end %end function interp1_scal

function jl = locate(xx,x)
%function jl = locate(xx,x)
%
% x is between xx(jl) and xx(jl+1)
%
% jl = 0 and jl = n means x is out of range
%
% xx is assumed to be monotone increasing

n = length(xx);
if x<xx(1)
jl = 0;
elseif x>xx(n)
jl = n;
else
jl = 1;
ju = n;
while (ju-jl>1)
jm = floor((ju+jl)/2);
if x>=xx(jm)
jl = jm;
else
ju=jm;
end
end
end

end %end function locate interp1, linear interpolation MATLAB Answers — New Questions

​

Hi,
For the example “Use FPGA I/O to Communicate with FPGA”, I followed all the documented steps exactly. I was able to generate the bitstream for the ZCU102 with FMCOM2/3 and successfully program the board using the generated bitstream.
Next, I copied the MATLAB code from the example page and ran it to transmit and receive a chirp signal using RF loopback. I can observe both the TX and RX power correctly. When I run the code a single time, it works as expected and I obtain the correct input and output signals.

However, I need to run this example inside a for-loop to transmit and receive data multiple times. I understand that on the rising edge of the "txStart" signal, TVALID pulses are generated on both the TX and RX sides. In the for-loop, the start signal is toggled by writing 1 and then 0, and the stream write and stream read operations are also executed.
I have tried different combinations of toggling the start signal and performing the stream read/write operations, but the behavior is not correct. Specifically, only one out of every three iterations receives the data correctly. you can find attached one of the multiple codes that I tried.

Could you please provide a MATLAB .m script that works correctly with this example when executed inside a for-loop? Are you able to send and receive data properly multiple times?
Thanks.Hi,
For the example “Use FPGA I/O to Communicate with FPGA”, I followed all the documented steps exactly. I was able to generate the bitstream for the ZCU102 with FMCOM2/3 and successfully program the board using the generated bitstream.
Next, I copied the MATLAB code from the example page and ran it to transmit and receive a chirp signal using RF loopback. I can observe both the TX and RX power correctly. When I run the code a single time, it works as expected and I obtain the correct input and output signals.

However, I need to run this example inside a for-loop to transmit and receive data multiple times. I understand that on the rising edge of the "txStart" signal, TVALID pulses are generated on both the TX and RX sides. In the for-loop, the start signal is toggled by writing 1 and then 0, and the stream write and stream read operations are also executed.
I have tried different combinations of toggling the start signal and performing the stream read/write operations, but the behavior is not correct. Specifically, only one out of every three iterations receives the data correctly. you can find attached one of the multiple codes that I tried.

Could you please provide a MATLAB .m script that works correctly with this example when executed inside a for-loop? Are you able to send and receive data properly multiple times?
Thanks. Hi,
For the example “Use FPGA I/O to Communicate with FPGA”, I followed all the documented steps exactly. I was able to generate the bitstream for the ZCU102 with FMCOM2/3 and successfully program the board using the generated bitstream.
Next, I copied the MATLAB code from the example page and ran it to transmit and receive a chirp signal using RF loopback. I can observe both the TX and RX power correctly. When I run the code a single time, it works as expected and I obtain the correct input and output signals.

However, I need to run this example inside a for-loop to transmit and receive data multiple times. I understand that on the rising edge of the "txStart" signal, TVALID pulses are generated on both the TX and RX sides. In the for-loop, the start signal is toggled by writing 1 and then 0, and the stream write and stream read operations are also executed.
I have tried different combinations of toggling the start signal and performing the stream read/write operations, but the behavior is not correct. Specifically, only one out of every three iterations receives the data correctly. you can find attached one of the multiple codes that I tried.

Could you please provide a MATLAB .m script that works correctly with this example when executed inside a for-loop? Are you able to send and receive data properly multiple times?
Thanks. hdl coder, zcu102, fmcomm2 MATLAB Answers — New Questions

​

Hello everyone,

I am simulating Orthogonal X-ray Projection, specifically Digitally Reconstructed Radiographs (DRR), from a particular CT volume.

To accomplish this, I first created a CT volume and then used the "generateOptimizedXrayProjections" function to produce DRR for both the anteroposterior (AP) and lateral projections.
1. To view the results from this process, please refer to lines 34-35 and line 53. When I run this simulation, I can see the DRR results.
2. However, when I used a real CT image (lines 38-39 and line 54), the generated DRR appeared completely dark.

My goal is to generate a DRR from a real CT image, such as CT_13.nii.gz.

I would greatly appreciate any insights or suggestions regarding this code.

Attached below are the MATLAB code (zip file) for reference.

Thank you!Hello everyone,

I am simulating Orthogonal X-ray Projection, specifically Digitally Reconstructed Radiographs (DRR), from a particular CT volume.

To accomplish this, I first created a CT volume and then used the "generateOptimizedXrayProjections" function to produce DRR for both the anteroposterior (AP) and lateral projections.
1. To view the results from this process, please refer to lines 34-35 and line 53. When I run this simulation, I can see the DRR results.
2. However, when I used a real CT image (lines 38-39 and line 54), the generated DRR appeared completely dark.

My goal is to generate a DRR from a real CT image, such as CT_13.nii.gz.

I would greatly appreciate any insights or suggestions regarding this code.

Attached below are the MATLAB code (zip file) for reference.

Thank you! Hello everyone,

I am simulating Orthogonal X-ray Projection, specifically Digitally Reconstructed Radiographs (DRR), from a particular CT volume.

To accomplish this, I first created a CT volume and then used the "generateOptimizedXrayProjections" function to produce DRR for both the anteroposterior (AP) and lateral projections.
1. To view the results from this process, please refer to lines 34-35 and line 53. When I run this simulation, I can see the DRR results.
2. However, when I used a real CT image (lines 38-39 and line 54), the generated DRR appeared completely dark.

My goal is to generate a DRR from a real CT image, such as CT_13.nii.gz.

I would greatly appreciate any insights or suggestions regarding this code.

Attached below are the MATLAB code (zip file) for reference.

Thank you! drr, digitally reconstructed radiographs, x-ray projection MATLAB Answers — New Questions

​

I want my font on my figues and tables to match the rest of what i am writing, so I ussually use the LaTex font in matlab when making figues.
In the start of the script I would write:
set(groot,’defaultAxesTickLabelInterpreter’,’latex’)
set(groot,’defaulttextinterpreter’,’latex’)
set(groot,’defaultLegendInterpreter’,’latex’)
However, since I hav updated my MatLab to R2023b, the font has changed.
I have attached eksample of how it looked before and after the update.

Granted symbols like and sub- and superscrpts still work as it ussually did.I want my font on my figues and tables to match the rest of what i am writing, so I ussually use the LaTex font in matlab when making figues.
In the start of the script I would write:
set(groot,’defaultAxesTickLabelInterpreter’,’latex’)
set(groot,’defaulttextinterpreter’,’latex’)
set(groot,’defaultLegendInterpreter’,’latex’)
However, since I hav updated my MatLab to R2023b, the font has changed.
I have attached eksample of how it looked before and after the update.

Granted symbols like and sub- and superscrpts still work as it ussually did. I want my font on my figues and tables to match the rest of what i am writing, so I ussually use the LaTex font in matlab when making figues.
In the start of the script I would write:
set(groot,’defaultAxesTickLabelInterpreter’,’latex’)
set(groot,’defaulttextinterpreter’,’latex’)
set(groot,’defaultLegendInterpreter’,’latex’)
However, since I hav updated my MatLab to R2023b, the font has changed.
I have attached eksample of how it looked before and after the update.

Granted symbols like and sub- and superscrpts still work as it ussually did. latex, font, r2023b MATLAB Answers — New Questions

​

Somehow I have reset Matlab to that I can only edit one script at a time. Can you suggest what I am doing wrong? Thanks.Somehow I have reset Matlab to that I can only edit one script at a time. Can you suggest what I am doing wrong? Thanks. Somehow I have reset Matlab to that I can only edit one script at a time. Can you suggest what I am doing wrong? Thanks. editor MATLAB Answers — New Questions

​

I am simulating a simplified lower-limb / prosthetic leg in Simscape Multibody with 3 revolute joints (hip, knee, ankle). The motion is intended to be planar in the sagittal plane (XZ).What I am doing
I provide reference joint trajectories for hip/knee/ankle from a gait dataset (10 concatenated gait cycles) using three timeseries signals:
hip_ts, knee_ts, ankle_ts
Each joint is controlled by a joint-level PID/PD block that outputs a torque
The measured joint angles q are taken from the joint sensors and fed back to the controller (scopes show reference vs simulated for each joint).
What works (no contact)
If I disable ground contact, the joint tracking is reasonably good for all joints (hip/knee/ankle). The simulated angles follow the reference signals with small error.
The problem (with contact enabled)
When I enable foot–ground contact using Spatial Contact Force (foot geometry vs ground plane), the tracking becomes much worse:
large tracking errors / oscillations appear
the behavior looks unstable or heavily disturbed during stance/contact phases
the same PID gains that work without contact no longer work
What I suspect / questions
I am trying to understand what is causing this degradation when contact is enabled:
Is this mainly due to the contact model parameters (stiffness, damping, friction) creating stiff dynamics that destabilize the joint PID?
Could the issue be caused by unintended extra DOFs at the hip/base (e.g., the hip not fully constrained in the plane), so contact forces introduce out-of-plane motion or additional dynamics?
Is it conceptually wrong to expect simple joint PID tracking to work well when contact forces are present, and should I instead use:
impedance control at the joints,
inverse dynamics / computed torque,
or a constraint-based approach for stance?
What I can provide
Screenshot of the Simulink/Simscape model (joint controllers + Spatial Contact Force block)
Scope screenshots showing reference vs actual joint angles for hip/knee/ankle
Contact block settings (stiffness/damping/friction) and solver settings if needed
Question: What are the most common reasons why enabling Spatial Contact Force makes joint-level PID tracking degrade dramatically in Simscape Multibody, and what specific checks/changes would you recommend (contact parameters, solver, constraints/DOFs at hip/base, control structure)?I am simulating a simplified lower-limb / prosthetic leg in Simscape Multibody with 3 revolute joints (hip, knee, ankle). The motion is intended to be planar in the sagittal plane (XZ).What I am doing
I provide reference joint trajectories for hip/knee/ankle from a gait dataset (10 concatenated gait cycles) using three timeseries signals:
hip_ts, knee_ts, ankle_ts
Each joint is controlled by a joint-level PID/PD block that outputs a torque
The measured joint angles q are taken from the joint sensors and fed back to the controller (scopes show reference vs simulated for each joint).
What works (no contact)
If I disable ground contact, the joint tracking is reasonably good for all joints (hip/knee/ankle). The simulated angles follow the reference signals with small error.
The problem (with contact enabled)
When I enable foot–ground contact using Spatial Contact Force (foot geometry vs ground plane), the tracking becomes much worse:
large tracking errors / oscillations appear
the behavior looks unstable or heavily disturbed during stance/contact phases
the same PID gains that work without contact no longer work
What I suspect / questions
I am trying to understand what is causing this degradation when contact is enabled:
Is this mainly due to the contact model parameters (stiffness, damping, friction) creating stiff dynamics that destabilize the joint PID?
Could the issue be caused by unintended extra DOFs at the hip/base (e.g., the hip not fully constrained in the plane), so contact forces introduce out-of-plane motion or additional dynamics?
Is it conceptually wrong to expect simple joint PID tracking to work well when contact forces are present, and should I instead use:
impedance control at the joints,
inverse dynamics / computed torque,
or a constraint-based approach for stance?
What I can provide
Screenshot of the Simulink/Simscape model (joint controllers + Spatial Contact Force block)
Scope screenshots showing reference vs actual joint angles for hip/knee/ankle
Contact block settings (stiffness/damping/friction) and solver settings if needed
Question: What are the most common reasons why enabling Spatial Contact Force makes joint-level PID tracking degrade dramatically in Simscape Multibody, and what specific checks/changes would you recommend (contact parameters, solver, constraints/DOFs at hip/base, control structure)? I am simulating a simplified lower-limb / prosthetic leg in Simscape Multibody with 3 revolute joints (hip, knee, ankle). The motion is intended to be planar in the sagittal plane (XZ).What I am doing
I provide reference joint trajectories for hip/knee/ankle from a gait dataset (10 concatenated gait cycles) using three timeseries signals:
hip_ts, knee_ts, ankle_ts
Each joint is controlled by a joint-level PID/PD block that outputs a torque
The measured joint angles q are taken from the joint sensors and fed back to the controller (scopes show reference vs simulated for each joint).
What works (no contact)
If I disable ground contact, the joint tracking is reasonably good for all joints (hip/knee/ankle). The simulated angles follow the reference signals with small error.
The problem (with contact enabled)
When I enable foot–ground contact using Spatial Contact Force (foot geometry vs ground plane), the tracking becomes much worse:
large tracking errors / oscillations appear
the behavior looks unstable or heavily disturbed during stance/contact phases
the same PID gains that work without contact no longer work
What I suspect / questions
I am trying to understand what is causing this degradation when contact is enabled:
Is this mainly due to the contact model parameters (stiffness, damping, friction) creating stiff dynamics that destabilize the joint PID?
Could the issue be caused by unintended extra DOFs at the hip/base (e.g., the hip not fully constrained in the plane), so contact forces introduce out-of-plane motion or additional dynamics?
Is it conceptually wrong to expect simple joint PID tracking to work well when contact forces are present, and should I instead use:
impedance control at the joints,
inverse dynamics / computed torque,
or a constraint-based approach for stance?
What I can provide
Screenshot of the Simulink/Simscape model (joint controllers + Spatial Contact Force block)
Scope screenshots showing reference vs actual joint angles for hip/knee/ankle
Contact block settings (stiffness/damping/friction) and solver settings if needed
Question: What are the most common reasons why enabling Spatial Contact Force makes joint-level PID tracking degrade dramatically in Simscape Multibody, and what specific checks/changes would you recommend (contact parameters, solver, constraints/DOFs at hip/base, control structure)? simulink, simscape, control MATLAB Answers — New Questions

​