But I can pass a reference to a matrix that will be used to update. In other words, I cannot write functions that return a matrix. RobotC does not support user defined data types as function types. The result of each function in the library is a scalar or matrix. This technique saves variable space and the amount of parameters used with functions. It is like me pointing out an article on Wikipedia instead of explaining things myself to you. Robotc supports “pass by reference” this means that one can tell a function were its input is located instead of passing al the input values.
#Robotc timer how to#
The third step is to decide how to pass information to my function library and back. All elements must be of type float because the filter uses real numbers, not integers. Here again, I sacrifice on flexibility and limit myself to 2×1 scalars and 2×2 matrices. The second step is to define the data types to store matrices and scalars. I will make a function library that works for these dimensions only. My model only uses scalars (2×1 matrices) and 2×2 matrices. This kind of flexibility is hard to get in robotC. That cannot be to hard, or can it? Well, matrices can have any number of rows and columns, a good function library should support this. There are only three matrix functions that I need: sum, product and inversion. The other option is to implement support for matrix functions. It is mainly for this reason I don’t go for this option. But more important it will make it hard to change the model If I need to. The disadvantage of this solution is that my functions can get complex and don’t resemble the model. The first option has the advantage of speed because it allows me to eliminate arguments and apply other optimizations. The first is to translate my filter into ordinary functions, the second is to write a function library to support matrix calculus. This leads me to the first decision I’ll have to make. Most computer languages don’t support matrices.
Robotc on the other hand has no native support for matrix language. They provide a convenient way to write down the various elements of the filter. The formulas of the Kalman filter use matrices from linear math. I use robotC as my programming environment but most things discussed in this post are relevant to all environments.
This time I’ll discuss the way I implement the filter on the NXT. If timer2 expires, make a small noise and reset it.In my previous post I discussed the Kalman filter for fusing compass and gyro readings. If timer1 expires, make a small noise and reset it.
#Robotc timer code#
* THIS CODE WILL ONLY WORK WITH ROBOTC VERSION 4.10 AND HIGHER * License: You may use this code as you wish, provided you give credit where its due. * The default number of timers is 10, but this can be changed by defining MAX_TIMERS * around 596 hours, it will roll over and weird stuff will happen. That means that if the program runs for more than
* timer.h provides additional timers for ROBOTC. Learn more about bidirectional Unicode characters To review, open the file in an editor that reveals hidden Unicode characters. This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below.
0 kommentar(er)
