Human Following Robot with implementation of FPGA`s

Since you are an aspiring Electronic Engineer I would expect you to approach this project the right way:

1) Specify what you want to do (the "requirements")

2) Work out how to do it in fairly abstract terms, which should give you an idea of the amount of processing power needed.

3) Work out what parts to use.

Because this is a student project you may need to modify the requirements spec (with the agreement of your tutor) to end up with a project which is feasible for you level of lnowledge and experience and the time available.

There is nothing in the basic requirement "Human Following Robot with Image Processing using FPGAs"

to suggest that an RPi or BeagleBone would be anything other than a distraction  - it sounds as if the emphasis of the project should be the FPGA not the supporting processor.

MK

Thank you Sir for your kind reply. Sir i have made my proposal and submitted it our group will be called to explain this project after that if the teachers agree only then will they give us permission to start this project. What i want to ask is that i have seen the Logi Pi, run some image processing algorithms to process images. Basically i want to divide the task the image processing part left to the FPGA whereas the motor, left right etc to the RPi. Yes Sir the emphasis is on the FPGA, but FPGA alone cannot do both tasks like image processing then controlling the motors etc, it does parallel processing that i know but image processing itself is a burdening process.

I think you would do far better to dump the pi and use an Arduino for the robot.

The FPGA in the LogiPi is a  Xilinx XC6SLX9-2TQG144  - it's a reasonable FPGA but you need to estimate the hardware requirements of the image processing algorithm BEFORE you choose the FPGA.

Steering the robot is likely to be trivial compared with the image processing.

It sounds to me as if you have very limited experience of FPGA use  - are you familiar with Verilog or VHDL (or Xilinix's own tools). Does your uni department have any tools available for you to use or expertise for you to call on for help (these factors should influence your choice of FPGA board and supplier, probably more than anything else.)

MK

Sir I do have a Arduino Mega 2560 with me. And I do have very limited knowledge of FPGA use, i dont have one, i have only read about it watched some videos, some projects based on FPGA`s. Sir but i am familiar with Verilog, the 6th semester which is finished now had a subject "Integrated Electronics" in which the whole semester we took labs and did that modelling, simulation etc. We used a software by SynaptiCAD named VeriLogger Pro. So Sir i do know VeriLog. Sir i have to search for some help in my uni dept. Sir my reason for selection of the package by ValentFX was its ease of use and we have a limited budget, if our project gets accepted the robot chassis, motors etc will be provided by my uni dept. but we will keep the enitre cost less than 2000$ thats like 0.2 Million in my country.

I agree with michaelkellett.

Using an Arduino would be an excellent approach.

Wanted to mention that this thread was mirrored here:

Human Following Robot with implementation of FPGA`s - Forum

If you have a limited budget you can use free tools from the FPGA vendor. You are likely to be choosing from Altera, Xilinx or Lattice. May advice is not to use kits from other parties that rely on their software and support. Using a third party board for the hardware is fine.

To do this project you need to start at the beginning and have some idea of the algorithms and signal processing techniques you will use. Then you can define the image processing hardware. You can simulate the signal processing in MATLAB and SIMULINK (which are available cheaply to you as a student) . You should do the simulation BEFORE you choose any hardware.

The control of the robot seems incidental to the project so use the Arduino and ready made (widely available) software for that.

If you have limited FPGA experience it is my honest opinion that this project is far too ambitious but if your uni insists you'll just have to do the best you can. Consider requiring the human to wear an easily identifiable marker (like a  target flag) .

How many people in your team and how many hours of work are expected ?

MK

Yes Sir I want to gather as much information as possible, so thats why i have been asking around i have posted on allaboutciruits.com, the valentfx forums & also over here at element14.

Thanks again Sir, Your advice to use Matlab & Simulink is really helpful, will do this before considering hardware. This project is really ambitious but if our proposal gets accepted, Sir we will try to complete it we will have a time of 1-year to complete it. We are 6 people, but 4 will work. Sir our teachers have said that we will consider you hard work even if your project doesn't work.

Thanks again Sir, Your advice to use Matlab & Simulink is really helpful, will do this before considering hardware. This project is really ambitious but if our proposal gets accepted, Sir we will try to complete it we will have a time of 1-year to complete it. We are 6 people, but 4 will work. Sir our teachers have said that we will consider you hard work even if your project doesn't work.

A team of six for  a year is enough effort to do some real work - at least one of you and probably three of you need to get good at FPGA stuff. A project of this size should not be looking at hobby stuff like the Logi (no slur on it intended - it's a nice enough product for hobby use and simple projects).

(The rest of the team needs to look at signal processing and hardware design.)

Once you get started it is very hard to reverse a decision on FPGA vendor - in your case there is little to choose between Altera, Xilinx and possibly Lattice. I would base my choice on the availability of expert guidance with the supplier in your uni. (And if the uni already has tools that can be important because free tools don't work for big FPGA projects).

You need to get proficient with HDL design and using the tools etc so once you've decided who to go with do some practice work with the tools. Now I see the size of the project I think you should control the robot with the FPGA - this is a relatively simple task and will help you get going with FPGA design and debugging before you get on to the hard stuff with image processing.

Because you've done  a course in Verilog that seems the obvious language to use but at least consider if you should go with VHDL (or possibly Xilinx HLS as part of their Vivado suite - very proprietary but might suit you and might be available free - ask Xilinx for support).

MK

Hello Sir I hope you are in fine and in best of health, tomorrow we are going to defend this project, I will tell them what I learned over here try to impress the teachers in order to get acceptance of this project, If it does get accepted I am going to be asking questions so I need your help.