![[PukiWiki]](image/pukiwiki.png "[PukiWiki]")
Conventional processors, like CPUs and GPUs, are efficient in processing dense, synchronously delivered data structures, not sparse, asynchronous event streams. Throughput is maintained by keeping the instruction and data pipelines as full as possible, even at the cost of executing redundant computations on unchanging data [Ref].
Recently, a new generation of natively event-based neuromorphic processors has appeared that can operate on sensor event streams directly [Ref.]. These many-core systems instantiate large populations of spiking neurons in massively parallel, low-power hardware, and inherit all of the advantages of event-based sensors, such as efficient, data-driven resource consumption[ Ref.].
Processing latency can be as fast as the event propagation time through the longest chain of neurons, so a neural network running on these systems can react to a stimulus in tens of milliseconds, fast enough to identify quick motions like hand gestures in real-time.
Brain-Computer Interfaces (BCI) are emerging devices that enable users to interact with computers by means of brain-activity only, this activity is generally measured by EEG sensors. BCI has various application, such as in the field of multimedia, virtual reality or video games. BCI also has other applications, such as for paralysed people, where it can be the only means of communication with the external world.
From another hand, the neuromorphic computing paradigm promises to drastically improve critical computational tasks' efficiency, such as decision-making and perception. Unlike the typical artificial neural networks (ANNs) where neurons fire at each propagation cycle, the neurons in a neuromorphic neural networks model, named spiking neural network (SNNs), fire only when a membrane potential reaches a specific value. Spiking neurons are only activated when sufficient signals are integrated from other neurons, which leads to sparse neural activities at the network level.
The goal of this thesis is to Study of EMG Signal Processing and Classification Algorithm Towards the Development of Adaptive Anthropomorphic Robot Arm.
| Date | Task |
| Oct 1 ~ Nov 30, 2020 | Complete running of ODIN for MNIST and Hand-gesture dataset |
| Oct 1 ~ Feb 28, 2021 | Survey of BCI, neuromorphic system and algorithms |
| Dec 1, 2020 | First draft submission deadline for ETLTC 2021 |
| Feb 28, 2021 | Final draft submission deadline for ETLTC 2021 |
| Mar 1 ~ Apr 30, 2021 | Propose new EMG and force classification algorithms |
| May 1 ~ June 30, 2021 | Implement new EMG and force classification algorithms |
| July 1 ~ Aug 31, 2021 | Evaluate new EMG and force classification algorithms |
| Sept 1 ~ Oct 31, 2021 | Create dataset |
| Nov 1 ~ Dec 31, 2021 | Integrate the developed algorithm with neuromorphic chip |
| Jan 1 ~ Feb 1, 2022 | Test the integrated neuromorphic chip for the dataset |
| Feb 2 ~ Mar 31, 2022 | Evaluate integrated neuromorphic chip |
| Apr 1 ~ Jul 31, 2022 | Write and submit a conference paper and journal paper |
| Aug 1 ~ Nov 30, 2022 | Study of different algorithms |
| Dec 1 ~ Dec 31, 2022 | Develop new signal processing algorithm |
| Jan 1 ~ Feb 1, 2023 | Write and submit a paper |
| Feb 2 ~ Feb 28, 2023 | Develop new classification and force algorithm |
| Mar 1 ~ Apr 30, 2023 | Integrate developed new algorithms with neuromorphic chip toward a neuromorphic arm |
| May 1 ~ May 31, 2023 | Evaluate the integrated neuromorphic chip |
| June 1 ~ July 31, 2023 | Write and submit a journal paper |
