The complexity of event-based object detection (OD) poses considerable challenges. Spiking Neural Networks (SNNs) show promising results and pave the way for efficient event-based OD. Despite this success, the path to efficient SNNs on embedded devices remains a challenge. This is due to the size of the networks required to accomplish the task and the ability of devices to take advantage of SNNs benefits. Even when "edge" devices are considered, they typically use embedded GPUs that consume tens of watts. In response to these challenges, our research introduces an embedded neuromorphic testbench that utilizes the SPiking Low-power Event-based ArchiTecture (SPLEAT) accelerator. Using an extended version of the Qualia framework, we can train, evaluate, quantize, and deploy spiking neural networks on an FPGA implementation of SPLEAT. We used this testbench to load a state-of-the-art SNN solution, estimate the performance loss associated with deploying the network on dedicated hardware, and run real-world event-based OD on neuromorphic hardware specifically designed for low-power spiking neural networks. Remarkably, our embedded spiking solution, which includes a model with 1.08 million parameters, operates efficiently with 490 mJ per prediction.

In the recent years, Spiking Neural Networks have gain much attention from the research community. They can now be trained using the powerful gradient descent and have drifted from the neuroscience to the Machine Learning community. An abundant literature shows that they can perform well on classical Artificial Intelligence tasks such as image or signal classification while consuming less energy than state-of-the-art models like Convolutional Neural Networks. Yet, there is very little work about their performance on unsupervised anomaly detection and time-series prediction. Indeed, the processing of such temporal data requires different encoding and decoding mechanisms and rises questions about their capacity to model a dynamical signal with long term temporal dependencies. In this paper, we propose for the first time a Sparse Recurrent Spiking Neural Network with specific encoding and decoding mechanisms to successfully predict time-series and do Unsupervised Anomaly Detection. We also provide a framework to describe in detail our model computational costs and fairly compare them with state-of-the-art models. Despite improvable performances, we show that our model perform well on these tasks and open a door for further studies of such applications for Spiking Neural Networks.

In this paper we propose an analytical Lithium Polymer (LiPo) rechargeable battery modeling approach to accurately predict battery SoC during simulations. This approach is based on a data-driven method for modeling the aging phenomenon present in rechargeable batteries. The proposed battery model is integrated into a system-level modeling methodology intended to simulate the discharge process of smart connected glasses powered by a 95mAh LiPo battery. Obtained results show that the proposed analytical battery modeling approach helps predicting battery SoC with less than 3% error. Moreover, this battery model allows simulating, in a few minutes, smart glasses scenarios that can last for days.

Bringing artificial intelligence to embedded devices has become a central research topic in many scientific domains (environment, agriculture, sociology, health...). For Human Activity Recognition, Artificial Neural Networks (ANNs) have shown their capability to provide better performance compared to other machine learning methods. However, ANNs suffer from two major limitations. First, ANNs are often trained using supervised learning requiring labelled databases, which are often difficult to build in real applications. Then, those algorithms are usually very expensive in terms of computing power. For that reason, their integration into low-power microcontrollers has been so far only evaluated to a limited extent. In this paper, we propose to evaluate quantitatively and qualitatively the embedded implementation of different neural networks for human activity recognition. First, supervised learning approaches are presented, followed by an exploratory study of unsupervised learning approaches using Self-Organizing Maps. Finally, some aspects of embedded unsupervised online learning are investigated to improve classification results using subject-specific data over a more general training. Each neural network is tested on a Human Activity Recognition dataset acquired from a smartphone using accelerometer and gyroscope sensing information (UCI HAR) and deployed on the SparkFun Edge board. This board hosts a low-power ARM Cortex-M4F-based microcontroller.

Wireless sensor networks (WSNs) with energy harvesting capabilities have drawn increasing attention in the last few years, as they enable long-term monitoring applications. However, the level of power harvested is usually limited to few mW. To improve the energy efficiency of WSNs, many power management techniques have been proposed to adjust the quality of service according to the harvested energy fluctuations. As wireless communications consume a major fraction of the available energy, numerous medium access control (MAC) protocols have been proposed to minimize energy consumption, latency, and data collisions. In this paper, we present an innovative MAC protocol for energy-harvesting based WSNs exploiting ultralow-power wake-up radios. To overcome the limited range typical of wake-up radios, a multi-hop wake-up scheme based on a dual radio system is proposed enabling asynchronous communications between a base station and any node of the network while maintaining a low latency and a high energy efficiency. To reduce energy consumption, wake-up calls and data packets are transmitted using two distinct data rates. Combined with destination address decoding, using a higher data rate for data transmission also minimizes the risk of collisions. Our approach has been applied to monitoring applications composed of autonomous sensor nodes powered by indoor light energy. OMNeT++ simulation results demonstrate the benefits of our wake-up radio-based approach in terms of energy, latency, and collisions when compared with the state-of-the-art duty-cycled MAC protocols. Experiments performed with real WSN platforms equipped with a wake-up radio prototype confirm the efficiency of our approach.

Wireless Sensor Networks (WSN) provide a powerful combination of distributed sensing, computing and wireless communication that can be useful in various applications (monitoring, health and fitness, environment, automotive, smart building, etc.). However, energy consumption is a critical issue in the deployment of battery-powered wireless sensor networks. Therefore, energy harvesting provides a potential solution to extend the system lifetime. This paper addresses the problematic of using wind as harvesting source, and weather forecast to improve the prediction of wind condition in the near future. The related issues concern the wind behavior in outdoor condition which is irregular and unpredictable. A wind harvester (a small wind turbine) is considered in order to define a model that translates a weather forecast (the wind speed) into a corresponding energy harvesting prediction. The proposed network architecture has therefore to manage the retrieval of weather forecast from Internet and transmit this information to the end devices. Moreover, the renewable energy storage device needs to be properly sized in order to provide power to the end devices when no more energy can be harvested from the ambient environment. For that, it is required to store a part of the harvested energy when it is available, and this is the role of a power manager (PM). In this paper, we propose different PM policies, two of them exploiting the use of weather forecast to improve the prediction accuracy for WSN nodes.

The paper focuses on techniques for the verification of software implementing communication functionality in networked embedded systems. We discuss the merits and limitations of tools for the simulation of a networked embedded system executing the binary code of the network protocol stack. In particular, we compare different solutions to extend a virtual platform to simulate the node of interest in a realistic communication scenario involving different network nodes. We then explain how this solution has the potentiality to perform verification of the protocol stack, which would be a great asset for industry and academia to validate the communication software under development or use.

Due to the ever-increasing demands on energy efficiency, designers are struggling to construct efficient and correct power management strategies for complex System-on- Chips (SoCs). The validation of an efficient power intent for a SoC is challenging and should be considered at early stage of the electronic system-level (ESL) design flow. To tackle this issue, we propose a high-level modeling approach on top of SystemC/TLM standard allowing the control structure of a power intent to be described in relationship with a functional transaction-level model (TLM) of a SoC. We consider a separation of concern approach in order to make easier the exploration of power intents and the optimization of power consumption. In this paper, the focus is set on an abstract clock intent model through a generic library and a modeling methodology allowing the development of a power management strategy on top of a functional SystemC-TLM model of a virtual prototype. A case study is used to demonstrate by simulation experiments the efficiency of this approach illustrating its capability to analyze effects of power management on performance and power consumption.

To design autonomous wireless sensor networks (WSNs) with a theoretical infinite lifetime, energy harvesting (EH) techniques have been recently considered as promising approaches. Ambient sources can provide everlasting additional energy for WSN nodes and exclude their dependence on battery. In this article, an efficient energy harvesting system which is compatible with various environmental sources, such as light, heat, or wind energy, is proposed. Our platform takes advantage of double-level capacitors not only to prolong system lifetime but also to enable robust booting from the exhausting energy of the system. Simulations and experiments show that our multiple-energy-sources converter (MESC) can achive booting time in order of seconds. Although capacitors have virtual recharge cycles, they suffer higher leakage compared to rechargeable batteries. Increasing their size can decrease the system performance due to leakage energy. Therefore, an energy-neutral design framework providing a methodology to determine the minimum size of those storage devices satisfying energy-neutral operation (ENO) and maximizing system quality-of-service (QoS) in EH nodes, when using a given energy source, is proposed. Experiments validating this framework are performed on a real WSN platform with both photovoltaic cells and thermal generators in an indoor environment. Moreover, simulations on OMNET++ show that the energy storage optimized from our design framework is utilized up to 93.86%.

In this paper, we propose a global power management approach for energy harvesting sensor nodes. Our approach is based on a joint duty-cycle optimization and transmission power control. By simultaneously adapting both parameters, the node can maximize the number of transmitted packets while respecting the limited and time-varying amount of available energy. We obtain a high-packet delivery by using an original predictive transmission power control that can efficiently adapt the transmission power to the wireless channel conditions. To accurately model the wireless channel and the node communication hardware, a waveform-level radio frequency simulator has been developed. Simulation results show 6.5 times improvement in energy efficiency and a packet reception ratio which is 9 times more efficient than a recently published technique. A 15% increase in energy efficiency, with respect to a fixed transmission power configuration, has also been observed. Finally, the global power management strategy has been validated on a real wireless sensor networks platform. Experimental results are very similar to those obtained in simulations, and thus confirm the efficiency of our power management approach.

Exploiting energy from the environment to extend the system lifetime of Wireless Sensor Network (WSN), especially thermal energy, is considered as a promising approach. When considering self-powered systems, the Power Manager (PM) plays an important role in energy harvesting WSNs. Instead of minimizing the consumed energy as in the case of battery-powered systems, it causes the harvesting node to converge to Energy Neutral Operation (ENO) in order to achieve a theoretically infinite lifetime. In this paper, a low complexity PM for a thermal-powered WSN is presented. Our PM adapts the duty cycle of the node according to the estimation of harvested energy and the consumed energy provided by a simple energy monitor for a super capacitor based WSN to achieve the ENO. Experiments are performed on a real WSN platform where harvested energy is extracted from the wasted heat of a PC adapter by two thermoelectric generators.

Systems such as wireless handsets include an integrated System-On-Chip (SoC) often based on a multi-core architecture, with at least one Digital Signal Processor (DSP) and one Micro Controller Unit (MCU) core. Validating the application and modem software that runs on these devices is a lengthy task. Moreover, the earlier software developers can access a realistic platform of the system, the earlier they can validate their software and ship a product. Synopsys DesignWareTMVirtual Platform is a product to enable pre-silicon software development, which allows software developers to develop and test software earlier, thus reducing time-to-market. In order to demonstrate virtual platform capabilities, OMAPV1030 layer-1 stand-alone modem software has been validated on the corresponding virtual platform. Using this methodology, a GSM call and both GRPS and EGPRS packet transfers have been performed purely in simulation. In addition, this paper describes how RF and audio data, to feed the virtual platform, have been collected.

Continuous advances in processor and ASIC technologies enable the integration of more and more complex embedded systems. Since their implementations generally require the use of heterogeneous resources (e.g., processor cores, ASICs) in one system with stringent design constraints, the importance of hardware/software codesign methodologies increases steadily. Interfacing heterogeneous hardware and software components together through a communication structure is particularly error prone and time consuming. Hence, on the basis of a generic architecute dedicated to telecommunication and multimedia applications, we proposed an extended communication synthesis method that provides characterization of communications and their implementation schemes in the target architecture. This method takes place after the partitioning and scheduling phase and may constitute the basis of a back-end of a codesign framework leading to HW/SW integration.

Generation of optimized DSP code from a high level language such as C is very time consuming since current DSP compilers are generally unable to produce efficient code.We present a software estimation methodology from a C description that helps for a rapid development of DSP applications.Our tool VESTIM provides both a performance evaluation for assembly code generated by the compiler and an estimation of an optimized assembly code.Blocks of applications G.721 and G.728 have been evaluated using VESTIM.Results show that estimations are very accurate and allow software development time to be significantly reduced.