Abstract: Wireless sensor networks (WSNs) are attracting increasing attention from academia and industry due to their promising application prospects. This paper introduces the architecture of WSN nodes, analyzes and compares typical hardware platforms currently used domestically and internationally, focuses on the advantages and disadvantages of commonly used processors, RF chips, power supplies, and sensors in WSN nodes, and compares in detail the wireless communication technologies currently applied to WSNs. Keywords: Wireless sensor network, hardware platform, low-power wireless communication Introduction A wireless sensor network (WSN) is a network composed of sensor nodes that can monitor, sense, and collect various information (such as light intensity, temperature, humidity, noise, and harmful gas concentration) of objects of interest to observers in the node deployment area in real time. This information is then processed and transmitted wirelessly to the observer via a wireless network. WSNs have broad application prospects in military reconnaissance, environmental monitoring, medical care, smart homes, industrial production control, and commerce. In a sensor network, sensor nodes function as end nodes and routers: on the one hand, they collect and process data; on the other hand, they fuse and route data, integrating the data they collect and the data received from other nodes, and forwarding the data to the gateway node. Gateway nodes are typically limited in number and their power can usually be replenished. Gateways usually communicate with the outside world using multiple methods (such as the Internet, satellite, or mobile communication networks). Sensor nodes, on the other hand, are numerous and typically powered by non-replenishable batteries. Once a sensor node's power is depleted, it can no longer perform data acquisition and routing functions, directly affecting the robustness and lifespan of the entire sensor network. Therefore, sensor networks primarily focus on sensor network nodes. While the design of sensor network nodes varies depending on the specific application, their basic structure remains the same. A sensor network node generally consists of four parts: a processor unit, a wireless transmission unit, a sensor unit, and a power module unit, as shown in Figure 1. Figure 1: Typical Components of a Wireless Sensor Network Node As a miniaturized embedded system, the sensor network node constitutes the foundational support platform of a wireless sensor network. Because wireless sensor networks are mostly battery-powered, operating in harsh environments, and are numerous, battery replacement is very difficult. Therefore, low power consumption is one of the most important design principles for wireless sensor networks. From the hardware design of wireless sensor network nodes to the protocol design of each layer of the entire network, energy saving is one of the design goals to extend the lifespan of the wireless sensor network as much as possible. Due to varying application contexts, a variety of hardware platforms for wireless sensor network nodes have emerged both domestically and internationally. Typical nodes include the Mica series, Sensoria WINS, Toles, μAMPS series, XYZnode, and Zabranet. The main differences between these platforms lie in the use of different processors, wireless communication protocols, and application-specific sensors. Commonly used wireless communication protocols include 802.11b, 802.15.4 (ZigBee), Bluetooth, UWB, and custom protocols; processors range from 4-bit microcontrollers to high-end 32-bit ARM core processors. Another type of node uses a microcontroller with an integrated wireless module, typically WiseNet. Typical wireless sensor network nodes are listed in Table 1. This paper introduces the concept and characteristics of wireless sensor networks and the composition of wireless sensor network nodes, focusing on analyzing and comparing the characteristics of various commonly used chips in each component unit, and consistently prioritizing low power consumption as a crucial comparison criterion. 2. Comparison of Typical Wireless Sensor Network Nodes Currently, researchers both domestically and internationally have developed various wireless sensor network nodes. While these nodes share similar components, their application backgrounds and performance requirements differ, leading to significant variations in the hardware components used. 2.1 Processor Unit The processor unit is the core of the sensor network node, working with other units to complete data acquisition, processing, and transmission/reception. The EM6603 is a 4-bit microcontroller with very low power consumption, but its processing power is also very limited. The Mica series nodes developed by the University of Berkeley mostly use microcontrollers from Atmel. Specifically, the Mica2 node uses the Atmel enhanced microcontroller ATmega128L. This microcontroller boasts abundant on-chip resources, including four timers, 4 KB SRAM, 128 KB Flash, and 4 KB EEPROM. It features UART, SPI, I2C, and JTAG interfaces for easy integration with wireless chips and sensors; it also offers six power-saving modes for convenient low-power design. Another advantage of using this processor is the availability of many compilers, including GCC (WINAVR), which is completely free and open-source software. Due to the advantages mentioned above and the influence of the Mica2 node, it is widely used in practical wireless sensor designs. However, from a low-power perspective, this chip is not the optimal choice. As listed in Table 1, in terms of low power consumption, the MSP430F1xx MCU series offers the industry's lowest current consumption, operating at 1.8 V, with a real-time clock standby current consumption of only 1.1 μA, and a running mode current as low as 300 μA (1 MHz). The entire wake-up process from sleep to normal operation takes only 6 μs. Low-power products have also been launched in the PIC series of microcontrollers. The Toles node and ZebraNet node use the MSP430 series microcontrollers, which have very low power consumption. In some applications with large data volumes, high-end processors are also used. For example, the μAMPS1 node uses the StrongARM processor SA1110, with a power consumption of 27–976 mW. This processor supports DVS power saving, which can reduce power consumption by about 450 mW; turning off the wireless module can reduce power consumption by 300 mW. The μAMPS2 uses a DSP processor. The XYZnode uses the OKI ML67Q5002 processor with an ARMTDMI core. This processor also supports DFS (Dynamic Frequency Scaling), with an operating current of 15–72 mA and a frequency of 1.8–57.6 MHz. Table 1 shows typical wireless sensor network nodes. From a processor perspective, wireless sensor network nodes can be broadly categorized into two types: One type uses high-end processors, represented by ARM processors. These nodes consume significantly more energy than those using microcontrollers, and most support energy-saving strategies such as DVS (Dynamic Voltage Scaling) or DFS (Dynamic Frequency Scaling). However, their processing power is also much stronger, making them suitable for applications with high data volumes, such as image processing. Furthermore, using high-end processors as gateway nodes is also a good choice. The last three processors in Table 2 are ARM core processors, whose power consumption is significantly higher than that of low-end microcontrollers. The other type consists of nodes using low-end microcontrollers. These nodes have weaker processing power but also lower power consumption. When selecting a processor, the system's processing power requirements should be considered first, followed by power consumption. Table 2. Performance Comparison of Various Common Microcontrollers 2.2 Wireless Transmission Technology and Chips Transmission media that can be utilized include air, infrared, laser, and ultrasound. Commonly used wireless communication technologies include: 802.11b, 802.15.4 (ZigBee), Bluetooth, UWB, RFID, IrDA, etc.; many chips also have user-defined communication protocols between the two parties. These chips generally operate in the ISM free frequency band, as listed in Table 3. Using laser as a transmission medium has lower power consumption and is safer than using electromagnetic waves. Disadvantages include: only straight-line transmission; susceptibility to atmospheric conditions; and directional transmission. These disadvantages determine that it is not an ideal transmission medium. Infrared transmission is also directional, has a short distance, and does not require an antenna. The 83F88S chip is a wireless transceiver chip conforming to the IrDA standard. UWB has advantages such as low transmitted signal power spectral density, low system complexity, insensitivity to channel fading, good security, high data transmission rate, and the ability to provide positioning accuracy of several centimeters; its disadvantage is that the transmission distance is only about 10 meters, and its penetration through walls is poor. 802.11b is not widely used due to its high power consumption. Bluetooth operates in the 2.4 GHz band, with a transmission rate of up to 10 Mbps; its disadvantages include a transmission distance of only about 10 meters, a complete protocol stack of 250 KB, unsuitability for low-end processors, and its primary use in home wireless LANs, though it also has some applications in wireless sensor networks. ZigBee and ordinary RF chips are the most widely used in wireless sensor networks. ZigBee is a short-range, low-complexity, low-power, low-data-rate, and low-cost bidirectional wireless communication technology. Its complete protocol stack is only 32 KB, can be embedded in various devices, and supports geolocation. These characteristics make ZigBee technology very suitable for wireless sensor networks. Currently, common chip manufacturers supporting the ZigBee protocol include Chipcon and Freescale Semiconductor. Figure8 has also developed a dedicated ZigBee protocol stack. Chipcon's CC2420 chip is widely used, with Toles and XYZ nodes employing this chip; Chipcon also provides a complete development kit including the ZigBee protocol developed by Figure8. Freescale Semiconductor offers ZigBee 2.4 GHz wireless transmission chips including the MC13191, MC13192, and MC13193; the company also provides accompanying development kits. Table 3 shows the wireless communication technologies used in wireless sensor networks. Ordinary RF chips are also an ideal choice, allowing for custom communication protocols. Representative MAC protocols include TMAC, SMA, CWiseMAC, BMAC, and DMAC. Routing protocols include Gossiping, SPIN, LEACH, and TEEN. Considering performance, cost, and power consumption, RFM's TR1000 and Chipcon's CC1000 are ideal choices. These two chips each have their advantages: the TR1000 has lower power consumption, while the CC1000 has higher sensitivity and a longer transmission distance. The WeC, Renee, and Mica nodes all use the TR1000 chip; Mica2 uses the CC1000 chip; Mica3 uses the Chipcon CC1020 chip, with a transmission rate of up to 153.6 kbps, supporting OOK, FSK, and GFSK modulation; and the Micaz node uses the CC2420 ZigBee chip. Another type of wireless chip integrates a processor, such as the CC2430, which integrates a 51-core microcontroller on top of the CC2420; and the CC1010, which integrates a 51-core microcontroller on top of the CC1000, further increasing the chip's integration. The WiseNet node uses the CC1010 chip. Other common wireless chips include Nordic's nRF905 and nRF2401 series, but due to their higher power consumption, lower receiver sensitivity, and greater development difficulty, they are less commonly used in practical wireless sensor networks. A comparison of the main parameters of commonly used wireless chips is shown in Table 4. Table 4 Comparison of Key Parameters of Commonly Used Wireless Chips 2.3 Power Module There are many types of batteries, and the energy storage capacity is related to factors such as shape, diffusion rate of mobile ions, and selection of electrode materials. Batteries in wireless sensor network nodes are generally not easily replaced, so battery selection is crucial, and the efficiency of the DC-DC module is also critical. Additionally, natural energy sources can be used to replenish battery power. Based on whether they are rechargeable, batteries can be divided into rechargeable and non-rechargeable batteries; based on electrode materials, batteries can be divided into nickel-cadmium batteries, nickel-zinc batteries, silver-zinc batteries, lithium batteries, lithium polymer batteries, etc. Generally, non-rechargeable batteries have higher energy density than rechargeable batteries; if there is no energy supply source, non-rechargeable batteries should be chosen. Among rechargeable batteries, lithium batteries and lithium polymer batteries have the highest energy density, but their cost is also relatively high; nickel-manganese batteries and lithium polymer batteries are the only non-toxic rechargeable batteries. Performance parameters of common batteries are listed in Table 5. Wireless sensor network nodes generally operate outdoors and can utilize natural energy sources to replenish battery power. Natural energy sources available include solar energy, electromagnetic energy, vibration energy, and nuclear energy. Because rechargeable batteries have a limited number of charge cycles and most have a memory effect, frequent charging using natural energy sources is not advisable, as it would significantly shorten battery life. Table 5 shows the performance parameters of common batteries. 2.4 Sensor Module Sensors come in many types and can detect physical quantities such as temperature, humidity, light, noise, vibration, magnetic field, and acceleration. Crossbow, Inc. in the United States has developed a series of sensor boards based on the Mica node, using sensors such as the Clairex CL94L photoresistor, the ERTJ1VR103J thermistor (Panasonic Electronics), the ADI ADXL202 accelerometer, and the Honeywell HMC1002 magnetic sensor. The SHTxx series temperature and humidity sensors support low-power mode, automatically entering sleep mode after data acquisition, with a current of less than 1 μA. The power supply circuit design of the sensor is crucial to the energy consumption of the sensor module. For sensors operating with low current (several hundred μA), they can be directly driven by the processor I/O port; when the sensor is not in use, the I/O port is set to input mode. In this way, the external sensor has no energy input and therefore no energy consumption. For example, the DS18B20 temperature sensor can use this method. For sensor modules operating with high current, the I/O ports cannot directly drive the sensor; MOSFETs (such as the Irlm16402) are typically used to control the energy input of subsequent circuits. When multiple high-current sensors are connected, integrated analog switching chips are usually used for power control; the MAX4678 is such a chip. 3. Conclusion Due to different application backgrounds, many hardware platforms exist both domestically and internationally, employing various wireless communication technologies. This paper mainly summarizes common wireless sensor network hardware platforms, analyzes and compares commonly used processors, wireless chips, wireless communication technologies, sensors, and power supplies, consistently considering power consumption as a crucial comparative factor. Through this detailed analysis of wireless sensor network hardware platforms, it is hoped that this will play a positive role in the research and development of wireless sensor networks in China.