2

I would like to do some customizations to IEEE 802.11 MAC layer protocol and implement it on hardware. I searched rigorously but I could not find any chipset with 802.11 PHY (RF + baseband chipset) using which I can implement custom MAC layer protocol. Most of the existing WiFi chipsets are integrated with MAC layer as its better and efficient to have it in hardware.

The one approach I think I have is to interface Radio frontend chipset (like MAX2830) with baseband chipset and then controlling everything using some MCU or FPGA. Is there any other easier approach I can use (e.g. SDR)?

Edit: The purpose I want to implement this is to achieve a custom wireless link with better packet priority and scheduling algorithm. I would also be interested in tweaking CSMA thresholds​ and implementing additional error correction coding in MAC layer itself.

Ashutosh
  • 798
  • 1
  • 5
  • 15
  • 4
    You’d be better finding a simulator that can do this. Approx. 10 years ago something similar was my final university project and I used OPNET. – David Jan 10 '18 at 12:09
  • Well, I want to improve quality of a physical wireless link and thus I will have to eventually implement it using hardware. – Ashutosh Jan 10 '18 at 12:32
  • 3
    *I want to improve quality of a physical wireless link* Then I think the MAC layer is a point which is too high, at the MAC layer the information is already in digital form so unless you add redundancy (so that data can be restored when parts are missing) your link will not improve. WiFi already does lower the datarate by adjusting the **Modulation** (for example use 16-bit QAM instead of 64-bit QAM) when reception is bad. Implementing your own modem (to do that modulation) for SDR is not a one-man-job, it requires a whole team and a lot of experience. I know, I worked on WiFi chips. – Bimpelrekkie Jan 10 '18 at 13:07
  • 1
    @Bimpelrekkie I apologise for not providing enough information. Plz see my edit. – Ashutosh Jan 10 '18 at 13:22
  • Why not use ns-3? It even allows you to interface with NI USRP so that the PHY remains in hardware, while all the desired functionality such as scheduling etc. can be coded. – V-Red Aug 15 '18 at 04:19
  • What are your statistics that warrant any changes and why do you feel your idea is better? You probably are getting Rician Fading – Tony Stewart EE75 Jan 22 '19 at 16:20

0 Answers0