Fwd: accelration/decelleration support

[Reza Naima]

I spent a couple hours simulating methods to do acceleration/deceleration yesterday, and I'm pretty close.  I've had almost success by linearly adjusting the time delay between steps downward to accelerate and upwards to decelerate.  I've written code in matlab to simulate and here's the results for moving to 7,20 from 0,0



The top two graphs are x and y vs time, and you can see the acceleration and deceleration.  the bottom graph is the position in x,y and it matches the expected slope perfectly.  there are a few more issues to work out, but as it stands, it should be easy to integrate.  It will be able to have an adjustable accel duration, decel duration.  Code below if anyone wants to play with it..

-reza


clear;
maxSpeed = 15;
stepsPerInch = 460;

start = [0 0 0];
finish= [7 20 0];
vector = finish - start;
steps = vector .* stepsPerInch;
location = start;

distance = sqrt( sum( vector.^2) );
duration = distance / maxSpeed;
timePerStep = duration ./ steps;
incTime = timePerStep;

accelTime=1;
decelTime=1;

s=9917
inclog = zeros(s,1);
t=zeros(s,1);
locx=zeros(s,1);
locy=zeros(s,1);
locz=zeros(s,1);
c=10;
scale=3/4;
finished = duration+accelTime*(scale)+decelTime*scale;
i=0;
for time=0:.0000001:finished
    for index=1:3
        if (steps(index)>0)
            if (time > incTime(index))
                i=i+1;
                if (time < accelTime)
                    inc = timePerStep(index) * (c-(c-1)*time/accelTime);
                elseif (time > finished-decelTime)
                    inc = timePerStep(index) * (1+(c-1)*(time-finished+decelTime)/decelTime);
                else
                    inc = timePerStep(index);
                end
               
                if (index==1)
                    inclog(i) = inc;
                end
               
                steps(index) = steps(index) - 1;
                location(index) = location(index) + (1/stepsPerInch);
                incTime(index) = time + inc;
                t(i)=time;
                locx(i) = location(1);
                locy(i) = location(2);
                locz(i) = location(3);
            end
           
        end
    end
   
end
%%
subplot(2,2,1)
plot(t,locx);
%axis([0 1 0 2.75])
title('X Axis vs. Time')
subplot(2,2,2);
plot(t,locy);
%axis([1 1.5 2.6 9])
title('Y Axis vs. Time')

slope = finish(2)/finish(1);
y=slope*locx;
err=abs(locy-y);

subplot(2,1,2);
plot(locx,locy,locx,err,'r',locx,y);
title('X Axis vs. Y Axis')

---

[iklln6]

logarithm that bitch

On Dec 26, 4:48 pm, Reza Naima <r...@reza.net> wrote:
> > I spent a couple hours simulating methods to do
> > acceleration/deceleration yesterday, and I'm pretty close.  I've had
> > almost success by linearly adjusting the time delay between steps
> > downward to accelerate and upwards to decelerate.  I've written code
> > in matlab to simulate and here's the results for moving to 7,20 from 0,0
>

> > ------------------------------------------------------------------------