Autonomous and Intelligent Wireless Transceivers

随着物联网概念的引入，无线设备和传感器的数量呈指数级增长，每三到四年翻一番。预计到2030年，这一数字将超过1250亿。然而，这种趋势给所有这些设备的维护、监视和控制带来了挑战。这些挑战激发了对自供电和自力更生的智能传感器和无线设备的需求。针对自主性，该研究计划提出了新的无线接收器架构和能量收集技术，以提高传感器和无线设备的能量效率。特别是，采用无源混合技术的高选择性接收器，将提出在给定载波频率下过滤接收到的期望信号，同时将反射的带内和带外干扰定向到能量收集电路。这些不需要的干扰中的能量将被利用并用于为接收器提供自供电。波形整形技术将用于构建高效、宽带和高动态范围的能量收集电路。在电路智能领域，减少对数字信号处理器传统确定架构的依赖，转向适合人工智能的神经形态架构，不仅可以显著降低信号处理组件所需的能量，还可以为这些传感器和设备带来更多的智能。这项研究计划将受益于人工智能算法和神经形态电路的进步。它将提出可重构和自适应模拟射频电路的新架构，以提高无线系统的智能。这些架构将通过整合自校准的简化神经形态电路元件，减少对人类输入的依赖。利用这一概念，我们将提出一种全数字自校准发射机。在这种智能发射机中，使用高效率的开关模式功率放大器将降低发射机的功耗。射频前端的自校准技术有望减轻这些非线性电路产生的失真，而不需要当前采用的解决方案中常用的数字预失真的计算复杂性。提出的新型智能自主发射器和接收器前端架构将成为未来几代无线通信标准中更具认知性、更普遍和更快的网络设计的使能技术。

With the introduction of the concept of Internet of Things, the number of wireless devices and sensors is exponentially increasing, doubling every three to four years. It is projected to reach over 125 billion devices by 2030. However, such trend creates challenges in maintaining, monitoring, and controlling all these devices. These challenges motivate the need for self-powered and self-reliant smart sensors and wireless devices. Aiming at autonomy, this research program proposes new wireless receiver architectures and energy harvesting techniques that improve the energy efficiency of sensors and wireless devices. In particular, a highly selective receiver using passive mixing techniques, will be proposed to filter through a received desired signal at a given carrier frequency, while directing the reflected in-band and out-of-band interferers toward an energy harvesting circuit. The energy in these unwanted interferers will be harnessed and used to self-power the receiver. Waveform shaping techniques will be used to build highly efficient, broadband, and high dynamic range energy harvesting circuits. In the area of circuit intelligence, less reliance on conventional definite architectures of digital signal processors and a shift toward neuromorphic architectures suitable for artificial intelligence will not only reduce significantly the energy needed by their signal processing components but also bring more intelligence to these sensors and devices. This research proposal will benefit from the advancement in artificial intelligence algorithms and neuromorphic circuitry. It will propose new architectures of reconfigurable and adaptive analog RF circuits that increase the wireless system intelligence. These architectures will rely less on human input by incorporating self-calibrating simplified neuromorphic circuit elements. Using this concept, an all-digital self-calibrating transmitter will be proposed. In this smart transmitter, the use of highly efficient switch mode power amplifiers will lower the power consumption of the transmitter. The Self calibration techniques of the RF front-end are anticipated to mitigate the distortion produced by these nonlinear circuits without the need for the computational complexity of digital predistortion commonly used in currently adopted solutions. The proposed new smart and autonomous transmitter and receiver frontend architectures will be an enabling technology for the design of more cognitive, more ubiquitous, and faster networks of the future generations of wireless communication standards.