Hi We have the position and velocity of products on the conveyor in time t=t1. How can we estimate position in time t=t1+T (T is scan time) if we know max.acceleration, max.jerk, max. velocity for this product? Product is on the conveyor. We need probably Kalman filter, but can somebody tell me formel how can we calculate position in t=t1+T. Thanks. Best regards Leo
Kalman filter estimation position
Started by ●October 27, 2008
Reply by ●October 27, 20082008-10-27
On Mon, 27 Oct 2008 00:26:46 -0700 (PDT), leo <e1e120032000@gmail.com> wrote:>Hi > >We have the position and velocity of products on the conveyor in time >t=t1. How can we estimate position in time t=t1+T (T is scan time) if >we know max.acceleration, max.jerk, max. velocity for this product? >Product is on the conveyor. We need probably Kalman filter, but can >somebody tell me formel how can we calculate position in t=t1+T.Maxima don't help. What you need is current estimates of position, velocity and maybe acceleration. If you have these, the estimated position at t2 is x1 + v1*T + a1T^2. Then you also update the estimates of x, v and a (and the state covariance estimate) by using the Kalman filter equations. Muzaffer Kal ASIC/FPGA Design Services DSPIA INC. http://www.dspia.com
Reply by ●October 27, 20082008-10-27
On Mon, 27 Oct 2008 00:26:46 -0700, leo wrote:> Hi > > We have the position and velocity of products on the conveyor in time > t=t1. How can we estimate position in time t=t1+T (T is scan time) if we > know max.acceleration, max.jerk, max. velocity for this product? Product > is on the conveyor. We need probably Kalman filter, but can somebody > tell me formel how can we calculate position in t=t1+T. > > Thanks. > > Best regards > > LeoIf _all_ you have is the position at time t = t1, then you can't do much. You need to have been reading the positions at sampling intervals all along; in that case you can build up an estimate of your state variables (which in this case seem to be position, velocity, and estimation). Unless your plant model and your disturbance models are known in great detail you may find that an H-infinity filter is better than a Kalman, although a Kalman may well be better than nothing. You'll also find that monitoring the drive command to the conveyor will be a help in accurate estimation, once again assuming that you have a good model. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●October 27, 20082008-10-27
Tim Wescott wrote:> (... position, velocity, and estimation).I take it that "position, velocity, and acceleration" was intended? ... Jerry -- Engineering is the art of making what you want from things you can get. ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by ●October 27, 20082008-10-27
Jerry Avins wrote:> Tim Wescott wrote: > >> (... position, velocity, and estimation). > > I take it that "position, velocity, and acceleration" was intended? >Yes. I think I'm a bit of a jerk this morning. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
