Battery-based sensor networks usually operate in harsh and dangerous remote environments that are inaccessible to people, such as volcanic zones and war zones. Power replacement or recharging of network nodes is usually impossible. The sensor nodes widely distributed in the tested environment must be responsible for collecting sensitive data, but also to complete the function of data transmission routing; moreover, an attacker may also use the encroaching node to inject a large number of false data packets into the network, causing the node to transmit These data packets run out of energy and lose their effectiveness. Therefore, the irreplaceability of the power supply of the network nodes makes the energy consumption problem particularly important compared to other key technologies of the sensor network; without affecting performance, designing effective energy consumption control strategies has become the core of the sensor network software and hardware design problem.
1 Sensor network node composition and energy analysis
A typical sensor network architecture is usually composed of distributed sensor nodes, receivers and transmitters, the Internet, and user interfaces. Among them, the sensor node as an independent working entity in the network, its basic functional subsystems include power supply subsystem, sensor subsystem, computing subsystem and communication subsystem, as shown in Figure 1.
1.1 Power supply subsystem
The power supply subsystem consists of modules such as batteries and ACDC converters. Its main task is to supply energy to other subsystems.
As the main energy source of the node, the battery's performance and capacity are very important. Although increasing the battery capacity can extend the energy supply time of the power supply subsystem, the use of effective recharging technology or renewable energy such as solar energy is more conducive to ensuring the energy source of the power supply subsystem and achieving continuous energy supply for other subsystems. A new i? Bean wireless transmitter based on i? Bean wireless technology and "energy acquisition" technology, which operates on an inductive oscillation energy converter [3], can be powered by 50 to 100 mg without battery power. The 28 to 30 Hz oscillation under force generates a voltage of 1.2 to 3.6 mV and allows data to be transmitted at a rate of 115 kbps over a distance of 30 m. Solution.
1.2 Sensor subsystem
The sensing subsystem is composed of a group of sensors and ADC controller, etc. The main task is to be responsible for sampling / collecting the sensitive information of the measured and controlled object and converting it into corresponding digital information.
Ideally, when the sensing subsystem automatically detects periodic and aperiodic events [4], its total energy consumption can be simply summarized as the product of the energy consumed by a single sampling and the number of samplings. Therefore, to control the energy consumption of the subsystem must be carried out from the following two aspects: one is to control the energy consumed by a single data sampling, and the other is to control the sampling frequency. The former can effectively control the energy consumption of single data sampling from the components themselves by using low power consumption devices. For the latter, since many distributed nodes in the sensor network are often groups of nodes to monitor the same object or sensitive data, selectively reducing the sampling frequency of a single node will not cause the validity and integrity of the measured data For destruction, as long as the activation principle of the node sampling task is properly set according to the application requirements, the energy consumption of the subsystem can be better controlled under the premise of ensuring data accuracy.
Figure 1 Block diagram of sensor network node structure
1.3 Computing subsystem
The computing subsystem includes microprocessors / microcontrollers, memory and I / O interface circuits and other hardware; it is responsible for controlling sensors, executing communication protocols and processing sensor data and other software algorithms; it is the node's control and computing core.
As the node's function control center and data computing center, the computing subsystem has complex functions and is closely connected with other subsystems. Therefore, the computing subsystem has strong and weak functions, high and low performance, and different working states (active, idle, and sleep, etc.) The duration of time and the switching between different states will seriously affect the energy consumption of the entire node. Low-power devices, timely sleep and idle frequency reduction techniques are all commonly used technologies to reduce the energy consumption of computing subsystems in hardware. The function rotation between nodes makes the energy consumption of network nodes relatively balanced from the overall network.
Self-organized cluster generation, transmission data encryption / decryption, and establishment and maintenance of communication links are all performed by executing corresponding instruction sequences. The more complex the algorithm, the more instructions, and the energy consumed The bigger. However, the algorithm is a contradictory unity of validity, reliability and complexity. Effective and reliable algorithms often have higher complexity; the effectiveness and reliability of simple algorithms may not be suitable for application needs. The diversity and uncertainty of the application environment make the energy consumption of software algorithms much more difficult to control than the energy consumption of hardware. It is necessary to meet the needs of the application environment and reduce the complexity of software algorithms as much as possible.
In addition, resource-constrained sensor network nodes are also vulnerable to physical damage attacks, making the control mechanisms and data processing algorithms commonly used in other computer networks, such as asymmetric key management protocols, unsuitable for sensor networks. According to the requirements of the application environment, sensor networks often have different levels of requirements for various control and data processing algorithms. Therefore, each control or data processing algorithm is a very challenging research field in sensor networks, and it is necessary to significantly modify the existing mature algorithms or redesign new processing algorithms according to the development level and technical characteristics of node energy. , Even when necessary; it can also control the energy consumption of nodes by appropriately reducing the performance of the network or nodes, in order to effectively extend the life cycle of the network.
Shenzhen Jiesai Electric Co.,Ltd , https://www.gootuenergy.com