Co-verification is the key step of software and hardware codesign on SOC. This paper presents a hw/sw co-verification methodology based on TWCNP-OS, a Linux-based operating system designed for FPGA-based platform of t...Co-verification is the key step of software and hardware codesign on SOC. This paper presents a hw/sw co-verification methodology based on TWCNP-OS, a Linux-based operating system designed for FPGA-based platform of two-way cable network (TWCNP) SOC. By implementing HAL (hardware Abstraction level) specially, which is the communications interface between hardware and software, we offer a homogeneous Linux interface for both software and hardware processes. Hardware processes inherit the same level of service from kernel, as typical Linux software processes by HAL. The familiar and language independent Linux kernel interface facilitates easy design reuse and rapid application development. The hw/sw Architecture of TWCNP and design flow of TWCNP-OS are presented on detail. A software and hardware co-verification method using TWCNP-OS is proposed, through the integrated using of Godson-I test board and TWCNP, which realizes the combination of design and verification. It is not a replacement of the co-verification with generic RTOS modeling, but is complementary to them. Performance analysis of our current implementation and our experience with developing this system based on TWCNP-OS will be presented. Most importantly, since the introduction of TWCNP-OS to our FPGA-based platform, we have observed increased productivity among high-level application developers who have little experience in FPGA application design.展开更多
提出一种基于SABO-GRU-Attention(subtraction average based optimizer-gate recurrent unitattention)的锂电池SOC(state of charge)估计方法。采用基于平均减法优化算法自适应更新GRU神经网络的超参数,融合SE(squeeze and excitation...提出一种基于SABO-GRU-Attention(subtraction average based optimizer-gate recurrent unitattention)的锂电池SOC(state of charge)估计方法。采用基于平均减法优化算法自适应更新GRU神经网络的超参数,融合SE(squeeze and excitation)注意力机制自适应分配各通道权重,提高学习效率。对马里兰大学电池数据集进行预处理,输入电压、电流参数,进行锂电池充放电仿真实验,并搭建锂电池荷电状态实验平台进行储能锂电池充放电实验。结果表明,提出的SOC神经网络估计模型明显优于LSTM、GRU以及PSO-GRU等模型,具有较高的估计精度与应用价值。展开更多
针对光伏发电应用领域太阳能路灯系统的过充电或过放电现象对蓄电池本身特性产生影响、降低使用寿命的问题,采用单片机和LabVIEW进行太阳能路灯蓄电池电压检测,采用BP神经网络进行太阳能路灯蓄电池荷电率(SOC)预测。BP神经网络将测得数...针对光伏发电应用领域太阳能路灯系统的过充电或过放电现象对蓄电池本身特性产生影响、降低使用寿命的问题,采用单片机和LabVIEW进行太阳能路灯蓄电池电压检测,采用BP神经网络进行太阳能路灯蓄电池荷电率(SOC)预测。BP神经网络将测得数据建立SOC(State of Charge)预测模型,LabVIEW可视化面板实时显示测量数据、波形及预测结果,实现太阳能路灯智能化控制。测试结果表明,系统能够实时检测蓄电池充电电压,并预测电池工作状态,BP神经网络蓄电池SOC预测值与蓄电池电量实测误差为0.1%~0.4%,满足网络误差要求。展开更多
文摘Co-verification is the key step of software and hardware codesign on SOC. This paper presents a hw/sw co-verification methodology based on TWCNP-OS, a Linux-based operating system designed for FPGA-based platform of two-way cable network (TWCNP) SOC. By implementing HAL (hardware Abstraction level) specially, which is the communications interface between hardware and software, we offer a homogeneous Linux interface for both software and hardware processes. Hardware processes inherit the same level of service from kernel, as typical Linux software processes by HAL. The familiar and language independent Linux kernel interface facilitates easy design reuse and rapid application development. The hw/sw Architecture of TWCNP and design flow of TWCNP-OS are presented on detail. A software and hardware co-verification method using TWCNP-OS is proposed, through the integrated using of Godson-I test board and TWCNP, which realizes the combination of design and verification. It is not a replacement of the co-verification with generic RTOS modeling, but is complementary to them. Performance analysis of our current implementation and our experience with developing this system based on TWCNP-OS will be presented. Most importantly, since the introduction of TWCNP-OS to our FPGA-based platform, we have observed increased productivity among high-level application developers who have little experience in FPGA application design.
文摘提出一种基于SABO-GRU-Attention(subtraction average based optimizer-gate recurrent unitattention)的锂电池SOC(state of charge)估计方法。采用基于平均减法优化算法自适应更新GRU神经网络的超参数,融合SE(squeeze and excitation)注意力机制自适应分配各通道权重,提高学习效率。对马里兰大学电池数据集进行预处理,输入电压、电流参数,进行锂电池充放电仿真实验,并搭建锂电池荷电状态实验平台进行储能锂电池充放电实验。结果表明,提出的SOC神经网络估计模型明显优于LSTM、GRU以及PSO-GRU等模型,具有较高的估计精度与应用价值。
文摘针对光伏发电应用领域太阳能路灯系统的过充电或过放电现象对蓄电池本身特性产生影响、降低使用寿命的问题,采用单片机和LabVIEW进行太阳能路灯蓄电池电压检测,采用BP神经网络进行太阳能路灯蓄电池荷电率(SOC)预测。BP神经网络将测得数据建立SOC(State of Charge)预测模型,LabVIEW可视化面板实时显示测量数据、波形及预测结果,实现太阳能路灯智能化控制。测试结果表明,系统能够实时检测蓄电池充电电压,并预测电池工作状态,BP神经网络蓄电池SOC预测值与蓄电池电量实测误差为0.1%~0.4%,满足网络误差要求。
