Co-design Scheme Research Articles

Software and hardware co-design and implementation of intelligent optimization algorithms

In order to improve the operating efficiency of the algorithm, some intelligent optimization algorithms are considered to be implemented on hardware. However, the existing design scheme has the problem of poor versatility. Therefore, this paper proposes a general software–hardware co-design scheme of intelligent optimization algorithms. In the design scheme, the initialization module and fitness module of the algorithm are deployed on the Advanced RISC Machines (ARM) for execution to increase the flexibility of the program. The update module of the algorithm is deployed on the Field Programmable Gate Array (FPGA) for execution to realize the hardware acceleration. The data between ARM and FPGA is transferred through Advanced eXtensible Interface (AXI) bus. In this paper, the PSO, BA, WOA, GWO, CMAES and EO algorithms are implemented with the proposed design scheme. And the six algorithms are tested on thirteen benchmark functions of different types. The experimental results prove the feasibility of the design scheme. In addition, by comparing with software and other implementation methods in execution time, resource occupancy and convergence, the effectiveness and superiority of the proposed scheme are proved.

Output Regulation State Bumpless Transfer Control for Switched Descriptor Systems

This brief is concerned with the state bumpless transfer control problem for switched descriptor systems with output regulation performance. Firstly, in order to alleviate state jumps at the switching instants while ensuring the solvability of the output regulation problem, a definition of the output regulation state bumpless transfer performance of switched descriptor systems is proposed for the first time. Then, sufficient conditions are put forward to ensure the output regulation performance and state bumpless transfer performance for the closed-loop switched descriptor systems based on the co-design scheme of switching rule and controller. Meanwhile, the co-design scheme gets rid of the requirement that each subsystem satisfies output regulation performance. In the end, a simulation example certifies that the given control technique is effective.

Non-fragile H ∞ control for event-triggered networked control systems with probabilistic time-varying delay

This paper studies the non-fragile output-feedback control problem for event-triggered networked control system with probabilistic time-varying delay and controller gain uncertainties. With known conditional probability distribution, the time-varying delay is modelled as a Bernoulli distributed white sequences. An improved two-side closed-loop functional is constructed by using the sawtooth structure characteristic of artificial time delay. By applying the Lyapunov–Krasovskii stability theory, the lower conservative stability criterions for time-delay system are derived. Aiming at the external disturbance and the parameter uncertainty of controller gain, a non-fragile output-feedback control strategy based on event-triggered mechanism is proposed. Moreover, a co-design scheme of controller gain and event-triggered matrix is obtained based on linear matrix inequality technique. Finally, three simulation examples are utilised to verify the effectiveness of the presented method.

Codesign of dynamic collision avoidance and trajectory tracking for autonomous surface vessels with nonlinear model predictive control

In this paper, a predictive control scheme for dynamic collision avoidance and formation trajectory tracking of autonomous surface vessels (ASVs) was designed. First, a trajectory-tracking nonlinear controller for ASVs was developed in the framework of model predictive control (MPC). Next, to realize leader-follower formation control, a graph topology approach was proposed to maintain a balanced distance between the leader and followers. Then, a set of nonlinear dynamic collision avoidance constraints based on the time to closest point of approach (TCPA) and distance to closest point of approach (DCPA) models was applied as control constraints. Specifically, the ship collision risk index (CRI) was proposed as a collision avoidance constraint for the controller. Simulation studies with various collision avoidance scenarios and predictive horizons were conducted using the dynamic CyberShip II model. The simulation results demonstrate the effectiveness of the proposed codesign scheme for formation tracking control and dynamic collision avoidance. Additionally, the simulation results show that the calculation time of the system was greatly reduced when using a trigger function, with an average calculation time of approximately 0.251 s.

Co-design of transition rates and sliding mode switched controller for Markovian jumping systems under intermittent DoS attacks

This paper focuses on the stabilization problem for a class of Markovian jumping systems (MJSs) subject to intermittent denial-of-service (IDoS) attacks by synthesizing the sliding mode control (SMC) and the transition rate matrix (TRM). The existing conditions for the transition rates are firstly established to ensure the exponential mean-square stability of the unforced uncertain MJSs. And then, a co-design scheme for both the sliding mode controller and TRM is synthesized to achieve the exponential mean-square stability of the closed-loop system under IDoS, in which a switching estimator is utilized to estimate the unmeasurable system state. By introducing a novel Lyapunov function, both the reachability and the stability of sliding mode dynamics are detailedly analyzed, and an iterative optimization algorithm is given for solving the corresponding sufficient conditions. Finally, the proposed co-design SMC strategy is illustrated via the simulation examples.

ReAAP: A Reconfigurable and Algorithm-Oriented Array Processor with Compiler-Architecture Co-Design

Parallelism and data reuse are the most critical issues for the design of hardware acceleration in a deep learning processor. Besides, abundant on-chip memories and precise data management are intrinsic design requirements because most of deep learning algorithms are data-driven and memory-bound. In this paper, we propose a compiler-architecture co-design scheme targeting a reconfigurable and algorithm-oriented array processor, named ReAAP. Given specific deep neural networks, the proposed co-design scheme is effective to perform parallelism and data reuse optimization on compute-intensive layers for guiding reconfigurable computing in hardware. Especially, the systemic optimization is performed in our proposed domain-specific compiler to deal with the intrinsic tensions between parallelism and data locality, for the purpose of automatically mapping diverse layer-level workloads onto our proposed reconfigurable array architecture. In this architecture, abundant on-chip memories are software-controlled and its massive data access is precisely handled by compiler-generated instructions. In our experiments, the ReAAP is implemented on an embedded FPGA platform. Experimental results demonstrate that our proposed co-design scheme is effective to integrate software flexibility with hardware parallelism for accelerating diverse deep learning workloads. As a whole system, ReAAP achieves a consistently high utilization of hardware resource for accelerating all the diverse compute-intensive layers in ResNet, MobileNet, and BERT.

Fault-tolerant control for linear parameter varying systems with integral measurements based on event-triggered mechanism

This study investigates the state and fault estimation, and fault-tolerant of discrete linear parameter varying (LPV) systems with integral measurements based on event-triggered mechanism (ETM). First, for LPV systems with integral measurements, a codesign scheme of a fault observer and a fault-tolerant controller based on ETM is proposed. Then, a sufficient condition for the proposed scheme is then specified and proved to meet the H∞ performance index. Upon, the gain matrices of the observer and fault-tolerant controller are obtained using linear matrix inequality (LMI). Finally, the feasibility and effectiveness of the proposed method are validated by a numerical example.

An Energy Efficient Computing-in-Memory Accelerator With 1T2R Cell and Fully Analog Processing for Edge AI Applications

In this work, a ReRAM-based energy-efficient CIM accelerator is presented with two techniques for edge AI applications. Firstly, a circuit-algorithm co-design scheme is proposed to realize fully analog processing, which improves the energy efficiency and the throughput of neural network. To deal with the I-V nonlinearity of ReRAM, a nonlinear-aware training algorithm is proposed to improve the network accuracy. Secondly, a 1T2R cell is proposed to replace previous 2T2R for weight storage with 35% area saving. For evaluation, a neural network with two fully connected layers and one ReLU layer is built for the MNIST dataset. The error rate can be reduced by >46% and the energy efficiency is 99 TOPS/W@200 MHz, 2.6X improvement over the digital method.

Event‐triggered adaptive tracking control of uncertain switched nonlinear systems

AbstractThis article is concerned with the event‐triggered adaptive tracking control of switched systems with uncertain nonlinearities and unknown control gains. The switch can occur within two consecutive sampling instants and may trigger new events that raise the mismatch problem and bring the challenge to avoid the Zeno behavior. To address these problems, we propose a novel codesign scheme composed of a switching event‐triggered mechanism and a mode‐dependent adaptive control law. A logarithmic threshold strategy is introduced into the event triggering mechanism. The absence of Zeno is proved from its formal definition. Based on the codesign scheme, the tracking error can be steered to an adjustable neighborhood of the origin. At last, a numerical example is applied to validate the obtained results.

A Co-Design Adaptive Defense Scheme With Bounded Security Damages Against Heartbleed-Like Attacks

This paper proposes a co-design adaptive defense scheme against a class of zero-day buffer over-read attacks that follow unknown stationary probability distributions. In particular, the co-design scheme integrates an improved UCB algorithm and a customized server. The improved UCB algorithm adaptively allocates guard pages on a heap based on induced damage of the guard pages so as to minimize the accumulated damage over time. The security damages of the improved UCB algorithm are proven to be always below a temporal bound without knowing which attack is launched when the buffer allocation follows a certain stationary probability distribution. Then an efficient server modification is introduced to randomly allocate buffers. Moreover, the damages of our scheme asymptotically converge to those of the optimal defense policy where the launched attacks and their distributions are known in advance. Further, the co-design scheme is evaluated with several real-world Heartbleed attacks. The experiment results demonstrate the validity of the upper bound and show that the adaptive defense is effective against all the attacks of interest with runtime overheads as low as 5%.

Model-Free Self-Triggered Control Co-Design for Probabilistic Boolean Control Networks

In this letter, a model-free co-design scheme of triggering-driven controller is proposed for probabilistic Boolean control networks (PBCNs) in order to achieve feedback stabilization with minimum controller efforts. Specifically, Q-learning (QL) algorithm is exploited to devise a self-triggered strategy wherein the controller update time is computed in advance by using the current state information. A new self-triggered QL (STQL) algorithm is presented to achieve the co-design of feedback controller and self-triggered scheme rendering the closed-loop system stable at a given equilibrium point. Finally, some examples are presented to demonstrate the effectiveness of the proposed method.

Reducing SRAM Reading Power With Column Data Segment and Weights Correlation Enhancement for CNN Processing

Convolutional neural network (CNN) has been widely deployed in various processors for intelligent visual signal processing. However, the large amount of activations and weights in CNN causes huge power consumption on SRAM access. Data-adaptive SRAM design is a widely studied method to reduce SRAM reading power based on the utilization of data patterns, while current designs only exploit data patterns in a coarse granularity, and have no advantages when faced with randomly distributed weight data. In this article, we propose a hardware–software co-design scheme to reduce SRAM reading power for CNN processing. First, we propose a reconfigurable data-adaptive SRAM architecture with column data segmentation (CDS-RSRAM) to utilize data patterns. Data in one column is partitioned into several segments, and finer-grained data patterns are exploited within each segment for further reading power reduction. Then, a novel training method—minimum segmented neighbor difference (miniSND)—is proposed for enhancing the correlation of weights. MiniSND improves the similarity of weights without classification accuracy degradation, thus weights could benefit from CDS-RSRAM and be read out with less power consumption. Simulation results demonstrate that the co-design scheme saves up to 66%(8b)/89%(2b) power consumption compared with 8T SRAM.

Model-dependent Scheduling and H-infinity Control Co-design for Networked Control Systems

In this paper, a novel model-dependent scheduling scheme is proposed for the networked control systems with time-delay, disturbance, and medium access constraints. The scheduler calculates the error between the ideal dynamic and the real system, and selects the states that make the stability of the system better to access the network. In addition, two kinds of representative time-delays in the networked control systems, constant time-delay and random time-delay, are considered. A robust H-infinity and switched-system-based co-design strategy is introduced to deal with the disturbance in the system, and the system uncertainty introduced by the random time-delay as well. Finally, illustrative examples are given to demonstrate the effectivity of the proposed scheduling method and the co-design scheme.

Event-triggered filtering for discrete-time Markovian jump systems with additive time-varying delays

This paper focus on the event-triggered filter design problem for delayed discrete-time Markovian jump systems. Firstly, a new mode-dependent Lyapunov-Krasovskii functional is introduced, where more than one Lyapunov matrices are dependent on jumping modes. Secondly, sufficient conditions are established by employing the Lyapunov stability theory together with Wirtinger-based inequality, under which the filtering error system is stochastically stable with strict dissipative. Thirdly, the co-design scheme of the filter parameters and event-triggered matrices is proposed. Finally, two numerical examples are given to show the usefulness of the proposed scheme.

Co-Design of Distributed Model-Based Control and Event-Triggering Scheme for Load Frequency Regulation in Smart Grids

In this paper, one new distributed load frequency regulation approach is proposed for smart power system operation under two specific practical constraints, including the limited communication resource and speed droop parametric uncertainty. To address these two constraints, the co-design of event-triggering communication scheme and distributed model-based controller is studied. Instead of using zero-order holders, the proposed model-based scheme is able to extend the maximum allowable time interval and thus reduce communication bandwidth usage. In the meantime, the proposed co-design scheme is able to get the model-based control parameters and event-triggering condition metrics simultaneously. This can loosen the conservation in the choice of control gains and event-triggering parameters faced by existing approaches where the control gains are fixed in prior. Comparisons on the multiple-area system confirm that this designed load frequency regulation method significantly reduces the number of required data transmissions without sacrificing the dynamic performance of the frequency and tie-line power. It is also shown that the proposed approach has great robustness to speed droop coefficient uncertainty.

NS-CIM: A Current-Mode Computation-in-Memory Architecture Enabling Near-Sensor Processing for Intelligent IoT Vision Nodes

In recent years, Neural networks (NNs) present vast potential for innovative applications. However, energy efficiency continues to remain a challenge in deploying NNs on the edge. In this context, computation-in-memory (CIM) architecture becomes an emerging trend in the area of energy-efficient hardware design, because it reduces data movement of multiply-accumulate (MAC) computation significantly. However, many recent works employ massive data converters to feed input data and transform output results, which may counteract the benefits of in-memory processing. To tackle this limitation, we propose a combined architecture cooperating sensor with CIM macro to achieve local processing of sensory signals. Current-mode computing techniques are exploited to achieve high energy efficiency while eliminating data conversion overhead. Moreover, we thoroughly analyze the non-idealities of the proposed mixed-signal circuits and present a co-design scheme to mitigate these imperfections. We have fabricated a 2K-bit CIM macro in the proposed architecture with TSMC 65-nm technology. The fabricated chip achieved 60.6 TOPS/W energy efficiency while consuming 845.5 μW power and 0.3 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> core area, presenting a promising solution for energy-constrained edge devices.

Dynamic Self-Triggered Controller Codesign for Markov Jump Systems

This article proposes a resource-aware triggering-based dynamic controller codesign scheme for Markov jump systems (MJSs) to fulfill both disturbance rejection and computational resources saving goals. Specifically, a self-trigger strategy is presented for precomputing the next execution time for state measurement, executing the controller, and updating the actuators using the current sampled information. To achieve a desired system performance and reduce the resource occupation, the self-triggering parameters and the gains of state feedback controller are codesigned. Moreover, a dynamic triggering technique is employed to improve the system performance by adapting the trigger coefficients to the different subsystems of MJSs. The effectiveness of the proposed method is demonstrated via two examples.

Dynamic Communication QoS Design for Real-Time Wireless Control Systems

In the coming fifth-generation (5G) cellular networks, ultra-reliable and low-latency communication (URLLC) is treated as an indispensable service to enable real-time wireless control systems. However, the extremely high quality-of-service (QoS) in URLLC causes significant wireless resource consumption. Moreover, to obtain good control performance may not always require extremely high communication QoS. In this paper, we propose a communication-control co-design scheme to reduce wireless resource consumption, where we obtain a dynamic communication QoS design method to reduce the energy consumption by jointly using extremely high QoS and a relatively low QoS. In this scheme, we first explore the control process served by different communication QoS levels and find that the whole control process can be divided into two phases, where different QoS levels have their advantages in different phases. Then, we obtain a threshold to decide when the extremely high QoS or relatively low QoS should be provided by communications. Simulation results demonstrate that our method can effectively reduce communication energy consumption while maintaining good control performance.

A Codesign Methodology for Constrained Networked Systems With Frequency Specifications

This brief addresses a controller/fault detector and scheduling protocol codesign scheme in the finite frequency domain for a class of constrained networked systems. By considering the data transmission in both sensor-control center and control center-actuator channels, a switched stochastic system is first modeled. With the aid of the mode-dependent average dwell time, sufficient conditions are subsequently investigated to capture the desired performances including the finite frequency one, in which the relationship between the node accessing rate and the performances is developed. Then, node-dependent controller/fault detectors and the scheduling protocol are codesigned by the derived conditions. Finally, an application to the High Alpha Research Vehicle (HASV) is given to illustrate the effectiveness of the proposed scheme.

Fixed-time sliding mode control based plant/controller co-design of dual-motor driving system

ABSTRACTA fixed-time sliding mode controller based plant/controller co-design scheme is developed for dual-motor driving systems. The proposed controller is fixed-time stable subject to unknown disturbances which guarantees the robustness of the proposed control scheme and a less conservative solution can be ensured by comparing with the conventional fixed-time convergence technique. Due to the existence of couplings between the plant and the controller, traditional design procedures in which the plant and the controller are designed sequentially and separately cannot guarantee the overall optimum of control systems. A co-design scheme based on the nested optimisation strategy is developed in this paper. With the nested optimisation strategy, the co-design solution can be achieved efficiently and reliably. A combined cost function considering both the plant and the controller is developed so that the largest load can be achieved without degrading the control performance. The fixed-time convergent controller is employed and considered into the co-design scheme. The gravitational search algorithm is adopted to optimise controller parameters in the co-design scheme. With the presented co-design method, the largest load of the studied system can be achieved with the satisfactory control performance. Contrastive Simulation and experimental results demonstrate the efficacy of the proposed controller and co-design scheme.