Course completion matrix

This matrix shows what each course block should contain before it can be considered complete. It is intended for course planning, GitHub issue tracking and quick repository review.

Legend

Mark	Meaning
✅	Complete and connected to the course
🟡	Partially complete or needs polishing
⬜	Planned / missing
N/A	Not applicable

Block-level matrix

Block	Theory	Python	MATLAB	C++	Fixed-point	HDL / FPGA	RF / measurement	Plots	Report task
01. Intro to SDR	✅	N/A	N/A	N/A	N/A	N/A	🟡	✅	🟡
02. Signals and sampling	🟡	🟡	🟡	⬜	⬜	⬜	🟡	🟡	⬜
03. DSP basics	🟡	🟡	🟡	🟡	🟡	🟡	⬜	🟡	🟡
04. Simulink and fixed-point	🟡	N/A	🟡	⬜	🟡	🟡	⬜	🟡	🟡
05. FPGA / HDL flow	🟡	🟡	⬜	⬜	🟡	✅	⬜	🟡	🟡
06. RF frontend and AD9363	🟡	🟡	⬜	⬜	N/A	⬜	🟡	🟡	🟡
07. TX/RX chains	🟡	🟡	⬜	⬜	🟡	⬜	🟡	🟡	🟡
08. Modulation and synchronization	🟡	🟡	🟡	⬜	🟡	⬜	⬜	🟡	🟡
09. Recording and analysis tools	🟡	✅	🟡	⬜	N/A	N/A	🟡	✅	🟡
10. KiCad and basic electronics	🟡	N/A	N/A	N/A	N/A	N/A	🟡	⬜	🟡
11. Integrated SDR project	🟡	✅	🟡	⬜	🟡	✅	🟡	🟡	✅
12. Final projects	🟡	N/A	N/A	N/A	N/A	🟡	🟡	⬜	🟡

Interpretation notes

• Block 11 is closed: the in-fabric QPSK modem runs over a two-board 915 MHz RF link. Lab 11.35's payload telemetry localized the single-bit misses to payload; Labs 11.41–11.45 then traced and removed the causes: the DC blocker tracking the modulation, a frame-sync false lock, and finally the four-branch rotation ambiguity itself, removed by differential coding plus a 24-symbol preamble. The final link has zero whole-burst rotation failures and payload BER ~4×10⁻⁴ (fabric loopback BER < 5.34×10⁻⁷ over 5.6 M bits); Lab 11.4 is the final measurement report. The residual differential single-bit rate is the inherent ~3 dB penalty, not an open rotation failure.
• Block 12 now has a filled dual-modem report with internal and cross-session external RF evidence. It remains below complete until controlled-path and longer-duration statistics are measured on a stable capture backend.
• Blocks 05 and 09 are the strongest automation anchors today: they combine reusable checks, generated artifacts and CI coverage.

Lab-level quality gates

Gate	Requirement	Evidence in repository
G1	Lab has a clear engineering objective	Lab markdown page
G2	Input and output signals are specified	Parameter table
G3	Reference model exists	Python and/or MATLAB script
G4	Figures are reproducible	Script + generated image
G5	Metrics are defined	SNR, EVM, BER, frequency error or implementation error
G6	Hardware relevance is stated	Fixed-point/FPGA/RF section
G7	Measurement workflow exists	IQ metadata file or capture plan
G8	Report task is included	Report checklist or template link

Recommended next content batches

Batch A: make Block 3 complete

• Finish FIR, mixer and decimation labs with Python + MATLAB + C++ reference code.
• Add fixed-point tables for coefficient width, input width and accumulator growth.
• Add FPGA mapping diagrams for FIR, NCO/mixer and decimator.
• Add IEEE-style comparison plots: floating point vs fixed point.

Batch B: connect RF hardware

• Keep the existing AD9363 frequency-plan, gain/bandwidth checklist and RTL-SDR capture guide synchronized.
• Run the safe cabled loopback and AD9363 gain/overload characterization.
• Publish or externally archive the measured QPSK raw WAV beside the existing manifest, plots and metrics.
• Extend the selected-frame RF result to per-burst success rates and BER/EVM distributions.

Batch C: final report layer

• Extend the filled implementation report with external burst success-rate evidence.
• Add a publication-cleared RF dataset with BER/EVM/SNR figures.
• Add screenshots of the finished measurement report and generated figures.

Batch D: make Block 11 reviewer-friendly

• Extend the selected-frame implementation report into a statistically repeatable external model-to-measurement proof.
• Keep the detailed bring-up logs as background, not as the main reviewer path.
• Add a one-page summary table: model, implementation, capture/evidence, metrics, conclusion and limitations.
• Challenge the selected routed bitstream timing margin across repeat builds or seeds.

Review rule

A course page should not be merged as complete if it only explains theory. Every important block should have at least one of these engineering anchors:

• executable model;
• generated figure;
• fixed-point analysis;
• HDL mapping;
• RF measurement plan;
• report task.