Pulse Signal Encoding With Hidden Variables for Low-Bandwidth BMI Links
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing brain-machine interface technologies face challenges in transmitting high-throughput pulse signals over low-bandwidth wireless channels, leading to issues with precision, real-time performance, and storage requirements, particularly in scenarios involving intrusive electroencephalogram signals.
Innovation Solution
An encoding method that compresses pulse signals into hidden variables using convolutional neural networks, followed by quantization and index encoding to generate a vector index, which is transmitted to a decoding apparatus for reconstruction, reducing data transmission volume.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If pulse signals are directly transmitted over wireless channels, then transmission accuracy is maintained, but data transmission volume is too large for low-bandwidth channels
Solution Approach 1:
The pulse signal is segmented into multiple components: binary code representing pulse presence, amplitude information, and temporal information. This segmentation allows selective transmission of critical components while discarding redundant information, reducing data volume while maintaining essential signal characteristics for accurate reconstruction.
Solution Approach 2:
The patent extracts only the essential features from the pulse signal for transmission: whether a pulse occurred, its amplitude level, and its temporal position. The original high-volume raw signal data is replaced by this extracted feature set, dramatically reducing transmission requirements while preserving the core information needed for signal reconstruction.
2Measurement precision
If high sampling rate and high throughput are used, then signal precision is improved, but power consumption and bandwidth requirements increase
Solution Approach 1:
The system dynamically adjusts the transmission strategy based on signal characteristics. Instead of continuously transmitting at high sampling rates, the system transmits only when pulses are detected, using variable coding lengths and rates adapted to the actual signal content, thereby reducing average power consumption while maintaining precision when needed.
Solution Approach 2:
The patent changes the representation parameters from continuous high-rate sampling to discrete pulse event coding. The signal is represented by parameters such as pulse count, amplitude levels, and time stamps rather than continuous waveform data, fundamentally changing how information is encoded and transmitted to reduce power and bandwidth requirements.
3Measurement precision
If high sampling rate and high throughput are used, then signal precision is improved, but wireless bandwidth requirements increase
Solution Approach 1:
The pulse signal is segmented into discrete event components (pulse detection, amplitude classification, temporal marking) rather than continuous data streams. This segmentation enables compression by transmitting only the essential event parameters, reducing the effective bandwidth required while maintaining signal precision through intelligent feature extraction.
Solution Approach 2:
Instead of copying the entire original signal waveform, the system creates a compressed representation copy that captures only the salient features: pulse occurrence times, amplitude levels, and pattern information. This copied representation is sufficient for accurate signal reconstruction at the receiving end but occupies far less bandwidth.
Data Source
AI summary
Embodiments of this application provide an encoding method, a decoding method, an encoding apparatus, and a decoding apparatus, and relate to the field of chip technologies. A method includes: obtaining a first pulse signal; encoding the first pulse signal to obtain a hidden variable; quantizing the hidden variable, and performing index encoding on a quantized hidden variable to obtain a vector index of the first pulse signal, where the vector index is used by a decoding apparatus to reconstruct the first pulse signal with reference to the hidden variable; and sending the vector index to the decoding apparatus.