Can u pls give some ideas to predict the trajectory of a moving object. I have planned to implement this trajectory prediction in hardware model too. Pls help me to do this . For this predicting process,which method i have choose,either neural network or some-else,,,.!!! ,pls help me with efficient method that i can implement in hardware too .
Without knowing more about your requirements, I think you should be able to do this with a microprocessor. Just define your equations for the trajectory, and put them into your code. What exactly do you mean by 'a hardware model'? Do you want to build an analog computer? (I don't think so). is your projectile just defined by an initial velocity and direction, or is it powered? Do you need to consider frictional forces? We need a little more information.
I think you might be able to do it with Kalman Filter, but that's all I know about that. I've never done this in pracitice, only learned theory in DSP courses though...
Without knowing more about your requirements, I think you should be able to do this with a microprocessor. Just define your equations for the trajectory, and put them into your code. What exactly do you mean by 'a hardware model'? Do you want to build an analog computer? (I don't think so). is your projectile just defined by an initial velocity and direction, or is it powered? Do you need to consider frictional forces? We need a little more information.
Thanks dude, My project is to predict the path of the flying object (unknown velocity). I have to do this prediction for a particular object in a video (which is a real-time input), I planned to place the video camera setup on a moving platform that will send the "live" video to the processing setup to make a trajectory prediction .
Kalman filtering sounds right (although I don't have any experience with that, either). Maybe you can do something like take two succesive samples, determine the velocity vector, and predict the position at the next sample instant. (I think this is sort of what Kalman filters do). Sorry I'm not much more help in this.
IIRC what Kalman filter does is taking few input samples including random noise etc, then try to predict value of the next sample. Next step is to compare the value predicted for particular sample with actual measured value of that sample. now from correlations between these two samples some corrections are made allowing for more accurate prediction.
i know people use those filters in miniature quadcopters and other flying RC models
Could the kalman filter technique applied for prediction too :!:. I came across that it can be used for tracking, but how about predicting the next instance of the object.
And also (plssss) assist me with step-wise procedure(algorithms,codes) to implement this prediction in both the real-time processor level and in matlab simulation.
kalman filtering can be used for prediction in a similar manner as other model-based methods. There is a prediction step in the kalman filter. The model-based methods provide some estimate of future values, and typically some bounds on the uncertainty. One of the difficulties is determining a suitable model for the system.
How to make use of kalman filter for prediction purpose.... can u pls give algorithm and coding for this prediction module... My project includes Video acqusition as real time input from a video camera. From this video sequence i have to do tracking of a particular object and also i have to do predict the trajectory of that particular projectile.
Which processor(peripherals) i have to choose ,.,.,.????