Can a pulse density modulated microphone be used in conjunction with a spiking neural network?
Best insight from top research papers
Yes, a pulse density modulated (PDM) microphone can be used in conjunction with a spiking neural network. The PDM microphone signal can be converted into rate coded spike frequencies, which can then be processed by the spiking neural network. This conversion from PDM to spike frequencies improves the time response of the neural network and allows for more accurate processing of the audio information. The conversion can be done using a spike-based system implemented in VHDL as an interface for PDM microphones, which converts the PDM pulses into temporal distributed spikes following a pulse frequency modulation (PFM) scheme with an accurate inter-spike-interval. This spike-based system has been tested and integrated with a neuromorphic auditory sensor (NAS) for verification .
Answers from top 4 papers
More filters
| Papers (4) | Insight |
|---|---|
Open access•Posted Content | Yes, the paper presents a spike-based system that converts audio information from pulse density modulated (PDM) microphones into rate coded spike frequencies for use in a Neuromorphic Auditory Sensor (NAS). |
Open access•Posted Content 30 Apr 2019 | Yes, the paper presents a sub-system that converts audio information from pulse density modulated (PDM) microphones into rate coded spike streams for a spiking neural network. |
| Yes, a pulse density modulated microphone can be used in conjunction with a spiking neural network. The paper mentions using PDM microphones to encode input sound into a spike train for processing with spiking neural networks. | |
12 May 2019 4 Citations | The answer to the query is not mentioned in the provided paper. The paper discusses the use of a two microphone system connected to a spiking neural network for acoustic source localization, but it does not mention the use of pulse density modulated microphones. |
Related Questions
Amplitude modulation in wireless microphone5 answersAmplitude modulation techniques have been proposed and investigated for various applications in wireless communications. In the context of indoor optical wireless communications, several modulation techniques have been compared and evaluated for their power and bandwidth efficiency. The Pulse Amplitude and Delay Modulation (PADM) technique has been proposed as a power efficient modulation technique, offering better bandwidth efficiency compared to other techniques such as PPM, DPPM, DPIM, and DH-PIM. Another modulation technique, Differential Amplitude Pulse-Position Modulation (DAPPM), has been shown to provide advantages over PPM, DPPM, and DH-PIMα in terms of bandwidth requirements, capacity, and peak-to-average power ratio (PAPR). Additionally, a tunable hybrid modulation technique has been proposed, which takes real-time channel conditions into account and aims to achieve a guaranteed system performance without compromising power and bandwidth efficiency. These modulation techniques offer insights into the power and bandwidth requirements of different amplitude modulation schemes in wireless communications.
Amplitude modulator in wireless microphone5 answersAmplitude modulation (AM) is a technique used in wireless communication systems, including wireless microphones. Several papers in the provided abstracts discuss the design and performance analysis of AM circuits. One paper proposes a modified low-power wideband on-chip AM based on independently driven double-gate metal–oxide–semiconductor field-effect transistor (IDDGMOS) devices, which shows higher gain and bandwidth for lower power input signals. Another paper presents a delta-sigma analog-to-digital converter (ADC) integrated with a bandgap voltage reference for direct voltage readout of an electret microphone, which operates at 64 times oversampling for a signal bandwidth of 11 kHz. Additionally, there are papers proposing power efficient modulation techniques for indoor optical wireless communications, which may have applications in wireless microphones. These papers discuss the advantages of different modulation schemes, such as pulse amplitude modulation (PAM) and pulse position modulation (PPM), in terms of power and bandwidth requirements.
What is the relationship between spiking neural networks and neuromorphics?5 answersSpiking neural networks (SNNs) are artificial networks inspired by the human brain that use neurons to perform computation. SNNs process binary and sparse spiking signals in an event-driven, online fashion. Neuromorphic computing platforms are emerging as energy-efficient co-processors for implementing SNNs. These platforms simulate the architecture and information processing of the brain, providing new insights for building artificial intelligence (AI) systems. The development of neuromorphic hardware systems, from artificial synapses and neurons to spike-based computing platforms, has made significant progress in overcoming the limitations of traditional computing architectures. SNNs are more biologically realistic than conventional artificial neural networks (ANNs) and are better suited for processing spatio-temporal data. While SNNs still lag behind ANNs in terms of accuracy, they require fewer operations and are promising candidates for AI applications.
What are the current challenges in developing spiking neural networks?5 answersThe current challenges in developing spiking neural networks include the need for automated tools and expertise in multiple domains, as well as the efficient partitioning and placement of large-scale computations on hardware. Additionally, the application of spiking neural networks in real-world and complicated machine learning tasks has been limited. However, recent studies have shown that spiking neural networks have great potential in solving such tasks due to their energy efficiency and temporal dynamics. Furthermore, the simple learning rules developed for spiking neural networks, such as STDP and R-STDP, can be a potential alternative to replace the backpropagation algorithm used in traditional neural networks.
How do spiking neural networks compare to other types of neural networks for prediction tasks?5 answersSpiking neural networks (SNNs) have emerged as a promising alternative to traditional deep neural networks for prediction tasks. SNNs offer advantages such as low-power computing, energy efficiency, and robustness to noise. However, the constraint of firing at most once, known as Time-To-First-Spike (TTFS), limits the capabilities of SNNs. To address this, a relaxed version of the Fast & Deep method has been proposed, allowing for multiple spikes per neuron. This relaxed constraint improves performance, convergence speed, sparsity, prediction latency, and robustness to noise compared to TTFS SNNs. SNNs also show potential for regression problems in engineering sciences, offering efficiency and precision while retaining generalizability. Additionally, delay-adaptive SNNs tailor the inference latency and energy consumption to the difficulty of each example, providing guaranteed reliability for decisions produced at input-dependent stopping times.
How to use spiking neural networks for neurological signals?5 answersSpiking neural networks (SNNs) can be used for neurological signals in various ways. One approach is to use SNNs for classifying input pattern signals. This involves implementing a network of spiking neurons connected by synaptic elements that can adjust the weights applied to the input signals. Another method is to use SNNs to generate Pulse Width Modulation (PWM) signals. This can be achieved by using the Izhikevich neuron model and a Field Programmable Gate Array (FPGA) platform to produce PWM signals based on spiking patterns. Additionally, SNNs can be used for signal processing by processing signals from one neuron using synapse weight parameters and inputting the processed signals to another neuron in the network. Furthermore, SNNs have been shown to be effective in classifying motor imagery tasks from EEG signals. These tasks can be recognized using SNNs with temporal information, exploiting the capabilities of the network. Finally, SNNs can be used to transmit signals between neurons in a hardware-implemented network by modulating spikes based on synaptic weights, allowing for a higher dynamic range in the synapse output.