geeanlooca/deepspace-turbo

Implementation of the CCSDS (Consultative Committee for Space Data Systems) standard for Turbo Codes

Quartz v4

“[One] who works with the door open gets all kinds of interruptions, but [they] also occasionally gets clues as to what the world is and what might be important.” — Richard Hamming

Quartz is a set of tools that helps you publish your digital garden and notes as a website for free. Quartz v4 features a from-the-ground rewrite focusing on end-user extensibility and ease-of-use.

If you are looking for Quartz v3, you can find it on the hugo branch.

🔗 Read the documentation and get started: https://quartz.jzhao.xyz/

Join the Discord Community

Sponsors

deepspace-turbo's Issues

Simplification for turbocodes

Consider storing outputs of convolutional encoders on a matrix, in order to be able to directly feed the BCJR decoders with the correct bit stream. This should also simplify the management of puncturing patterns.

As a sidenote, there's the need to create a new function that, given sigma, adds noise to the transmitted packet.

Trellis termination is not exactly as in the standard

As of now, trellis termination generates all zeros in the systematic output, while in the proposed scheme the systematic output is equal to the feedback part. This is due to the fact that in the standard the systematic output is connected before the XOR block summing input and feedback.

Any way to evaluate the quality of the data block?

Hi Gianluna,

Thank you very much for your excellent work!

I am Wei Mingchuan from Harbin Institute of Technology. I am working on a lunar formation satellite project which will launch 2 micro-satellites to lunar helical orbit later this year. I build a GNU Radio OOT module for the downlink reception of the satellites, with your codes for turbo: https://github.com/bg2bhc/gr-dslwp/

During my recent tests I found that at low SNR sometimes the PLL of the receiver may faile before finished receiving a data block, this means the left half would be very unreliable. At this time turbo decoder will try to fix the errors but generate more errors. Is there any way to evaluate the quality of the data block?

Thank you very much!

Best Regards,

Wei Mingchuan

a_priori variable passed to convcode_extrinsic should be packet_length + code.memory long instead of packet_length

the function should be able to accept a packet_length + code.memory long vector for a-priori probabilities since this is necessary when decoding turbo codes, where Extrinsic messages from a different code are passed to the convcode_extrinsic function.

Memory allocation

Optimize memory allocation by pre-allocating the necessary space when initializing the code instead of allocating/deallocating inside each function at each call.

Possible bug when decoding a turbo code on multi-core system

Upon calling turbo_initialize() a memory block used to store interleaved messages is allocated.
If the function turbo_encode() is called inside a parallelized for loop, the same memory block is accessed by different threads, possibly corrupting the messages.