The system reachability set is calculated by covering all possible behaviours of the system through a finite number of simulation steps to ensure that the system trajectory stays within a set safety region. In this paper, the theory of the game method is applied to the design of the controller, a very small controller is designed, and good control results are obtained by simulation. The system gradually shows a divergent trend and cannot achieve stable control. A multihop channel reservation Medium Access Control (MAC) protocol based on a parallel mechanism is proposed. The multihop channel reservation mechanism is proposed based on periodic node sleep, and the node uses the reservation frame to make the reservation of the channel and transmits the data according to the reservation information; the parallel mechanism is used to make the reservation information and data transmission simultaneously.
Cyber-physical system (CPS) is a high-efficiency cybernetically networked intelligent information system that integrates computation, network, and physical entities based on environment sensing and integrates the 3Cs (computation, communication, and physical information) [
CPS Week is the most important event on information-physical fusion systems, with five leading conferences from all over the world covering all aspects of CPS, bringing together academics from all over the world to discuss technical developments [
CPS is a network with control properties, but it is different from the existing control systems. Information-physical converged systems are prone to failures, such as attacks on their physical infrastructure, as well as attacks on the data management and communication layers. Concerns about the security of control systems are not nascent, as there is much literature on system failure detection, isolation, and recovery. For some of the deficiencies that exist in CPS that do not affect the control system, appropriate detection and identification techniques need to be developed. The reliance on communication networks and standard communication protocols to transmit measurement and control packets increases the likelihood of attacks on physical systems. Information security methods, such as authentication, access control, and message integrity, do not appear to be sufficiently effective in protecting information physically fused systems. These security methods do not take advantage of measurements on controlled systems under physical processes or control mechanisms, and they are ineffective against attacks on physical systems. Information-physical fusion systems consist of deeply integrated, tightly coupled computational and physical components with communication capabilities. These systems are highly complex and span multiple scientific and technical domains. As a result, they also pose many challenges, and they are extremely important for the progress and security of society. PS is becoming more complex, open, and vulnerable to security threats, making security an important issue to consider when designing CPS.
CPS shares some common features with popular Information and Communication Technology (ICT) systems, such as embedded systems, Networked Control Systems (NCSs), the Internet of Things (IoT), and the Internet of the Future [
On the other hand, it emphasizes dynamic interaction between physical processes and networks. IoT is more like a platform for implementing several applications. In other words, it is an extension of the Internet. In contrast to the Internet of Things, CPS is a way to go about realizing and designing the real world. The coming Industrial Internet refers to the integration of global industrial ecosystems, pervasive sensing, advanced computing, and pan-network connectivity to the increasing benefits of the world economy. Thus, its technological foundation is CPS. Wireless sensor networks (WSNs) are one of the main ways of information access in CPS [
CPS framework.
Rapid advances have been made, but there will still be significant challenges in designing a CPS using existing underlying theory and techniques. The goal of a CPS is to achieve a real-time, efficient interaction between computational and physical processes. This close interaction, while having many advantages, also poses many risks. New types of risks include adverse effects of the information system on the physical environment as well as adverse effects of the physical environment on the information system. Therefore, security, reliability, defensiveness, confidentiality, and adaptiveness, which are less important features of traditional computing systems, are the focus of CPS. Research in related security areas has focused on the development of tools and techniques to address known security issues. Important security tools, such as antivirus software and intrusion detection systems (IDS), are important. This approach to security is effective in responding to attacks using off-the-shelf tools and popular techniques. However, the response measures to new attack vectors and methods are inadequate. This continues to push the security frontier and stimulate the need for proactive security approaches. An important feature of CPS is the seamless and complex interaction between the computing unit and the physical environment. CPS should be designed and implemented with strict security, which requires a safety verification of failures. In the field of safety-critical systems, safety verification of the system model using formal verification theories and techniques that can prove the correctness of the system model is extremely important and has been a major concern in recent times.
Safety in CPS can be divided into fault safety and active safety. Failure safety refers to the prevention of system failures caused by unintentional actions in normal operation and is the avoidance of occasional failures, while active safety prevents the loss of system functions caused by intentional damage by operators and focuses on the active prevention of malicious attacks. For the safety design of the CPS system, the main concern is how to prevent accidents, so safety research should be divided into the following two aspects: for the absence of an attacker, to ensure fault safety, the system’s reach should be calculated and verified; for the presence of a malicious attacker, to ensure active safety, the system should be resilient and have robust control. The safety test is to verify that, from the given initial conditions, the system will operate in violation of the statute or is unsafe. Due to the continuous advances in technology and the increasing complexity of technical systems, strong verification methods already exist for purely discrete systems that can prove the safe operation of quite complex systems, and for hybrid systems with discrete and continuous dynamic combinations, safety assessment is a common problem for many classes of systems. A simple way to check the safe operation of mixed systems is to apply Monte Carlo simulation. Monte Carlo simulation focuses on calculating the probability of system failure, and there are methods for generating test cases in such a way that makes it possible to maximize test coverage to detect unsafe executions [
Data fusion is a multilevel processing process that automatically detects, estimates, federates, correlates, and combines data from multiple sources. The working environment and node characteristics of WSNs determine the necessity of data fusion. Firstly, WSN nodes are energy-intensive and need to minimize the amount of data transmission to save energy. Sensor nodes are battery-powered, and most of them are manually scattered in random areas, making it difficult to replenish energy by replacing batteries. Once the network runs out of energy, the data transmission capabilities of the WSN will be lost. The main energy-consuming modules of the node are the wireless transmission module, sensor module, and processor module; with the support of advanced circuit technology, the energy consumption of the processor and sensor is very low, and the main energy consumption comes from the wireless AC module. Therefore, to extend the node life, it is highly necessary to fuse the data to reduce the wireless transmission capacity. Second, the data in WSNs is often redundant and needs to be fused to obtain more accurate information. Compared to transmitting aggregated data, transmitting raw data does not bring more information to users and may lead to data conflicts, and the retransmission mechanism in WSNs will further increase network energy consumption and reduce information collection efficiency [
WSNs data topology control model.
Among them, the probabilistic model-based algorithm is suitable for situations where the user needs to know not only the monitoring results but also all the information of each node, and the data transmission volume is huge; this type of algorithm does not contain an anomalous data processing module and the data within the tolerance range is not transmitted, so the accuracy is the lowest; moreover, the energy-saving of this type of algorithm depends on the specific setting of the model and tolerance level. Neural network algorithms have good energy savings and very low information transfer, but not high accuracy. The Kalman filtering algorithm reduces the effect of white Gaussian noise and thus has higher data accuracy than other algorithms; and it has good energy savings and very low data transmission and is widely applicable to many agricultural IoT monitoring systems.
Scale-free networks consider two important characteristics of real networks: the size of the network is not fixed, but is constantly expanding. New nodes joining the network are more likely to be connected to nodes of larger size [
The algorithm calculates the edge capacity based on the remaining energy of the nodes and periodically invokes the maximum flow algorithm to calculate the broadening path and traffic flow to maximize the network load traffic [
The edge capacity on the virtual source and virtual destination nodes and their neighbour nodes is determined by the ability of their neighbour nodes to forward packets. This is because the virtual source node and virtual destination node have infinite energy and the edge capacity is determined by their neighbour nodes with finite energy.
Preference is given to a node to send packets to its lower-level neighbour nodes for the sake of a minimum number of hops; a node is considered to send packets to its fellow neighbour nodes only if it does not have a lower-level neighbour node. The scale factor is the meaning of the percentage of the energy of the neighbour node R in the sum of the energy of all neighbour nodes. The average energy collected by the node in a second is
The elements of
For the Pierre Mendes France (PMFR) algorithm, if the source node sends packets at time
Similarly, it is possible to calculate the edge capacity now of C according to the following equations:
If it is the first packet transmission, there is no need to update the remaining energy of the node because the node has not started collecting energy; if it is not the first packet transmission, there is a need to update the remaining energy of the node [
When designing a MAC protocol, the actual requirements and overall performance of the network should be taken into consideration. Nodes receive and process unnecessary data from neighboring nodes. The crosstalk phenomenon causes the wireless transceiver and processing modules of nodes to consume more energy because of the design of the MAC protocol. In the specific application of wireless sensor networks, the number of sensor nodes and their distribution density in the network are subject to change. The death of old nodes as well as the addition of new nodes must require good dynamics of the network topology. Wireless sensor networks, as a distributed self-organizing network, require MAC protocols that are adaptable and scalable to cope with changes in network load and topology.
If there are too many control messages, the sensor network will also consume more energy to execute the control messages. However, there are specific applications where energy efficiency is less important than the packet-to-packet ratio and low latency, where high transmission rates can be achieved at the expense of energy efficiency. Communication protocols generally prioritize energy efficiency and are designed to tolerate latency and other issues. However, when it comes to specific application requirements, energy efficiency is less important than packet-to-packet ratios and low latency, and communication protocols should be responsive. The protocol must also be able to prioritize packets carrying critical data, as they often contain important sense data and require efficient transmission. Also, the protocol must support fairness to the source of the packet. Fairness is important when there is a hazard, and an aggregating node can receive complete information from all sensor nodes to monitor the propagation of the hazard.
The parent node responds to an acknowledgment (ACK) message to its children to confirm that each child node is added to its list of children. If a node does not receive an acknowledgment message after a specific period broadcast, the message is replayed. A node keeps broadcasting the construction tree information until it receives an acknowledgment message or exceeds the maximum value to retransmit. If a node updates its Parentnode_id and its Hop, it also needs to notify its old parent node after reestablishing the build tree information, and the old parent node responds to the node with an acknowledgment message informing the node that it has been removed from the list of child nodes. If the node does not receive the old parent’s acknowledgment message, it will rebroadcast the build tree information, and the old parent’s acknowledgment message will help the child node list the most recent information. If the list of child nodes is not updated, the old parent node may try to receive some packets from the child node, thus wasting energy on idle listening. The node will keep broadcasting the tree building information until it receives an acknowledgment or exceeds the maximum number of retransmissions. During the tree network building phase, the node may crosstalk to hear transmissions from other nodes within its transmission range. The node records the sender of the message as one of the nodes in its list of single-hop neighbour nodes. At the end of this phase, all nodes in the network have received the tree building message. When this phase is over, each node knows the number of hops that reached the convergence node, its parent, the list of children, and the list of single-hop neighbours’ nodes.
During this phase, nodes perform slot allocation and exchange scheduling plans, making sure that no node shares a slot between two hops, as their transmissions may collide if nodes have the same slot within two hops. At the end of the slot allocation phase, each node ensures that its scheduling plan is different from that of its single-hop and two-hop neighbours to avoid conflicts. Time slot allocation follows a bottom-up approach, starting with a leaf node (with no children), and the purpose of the time slot allocation job is to ensure that the transmission plan from the leaf node can support the flow of messages to the aggregating node. During the time slot allocation phase, all communications are scheduled for conflict-free periods using CSMA/CA.
In the LEACH algorithm, the setting of the number of cluster headers is an important factor that affects the energy consumption of WSN. If the number of cluster headers is too few, some nodes in the network are too far away from the cluster head locations, resulting in higher energy consumption for transmission between the nodes and the cluster headers; if the number of cluster headers is too high, the energy consumption for communication between the cluster headers and the base station is too high. Therefore, there should exist an optimal number of cluster heads that minimize the total energy consumption of the entire network. Figure
MAC protocol design.
Relationship between system energy consumption and percentage of cluster headers.
The energy consumption for node data transmission is closely related to the transmission distance, so the difference in network diameter will have an important impact on the energy consumption of the WSN. Figure
Total system energy consumption versus network diameter.
Figure
Comparison of energy consumption when network diameter and node energy change.
The trend of energy consumed as the packets sent varies from 2 bits to 4096 bits. The Genomics Education and Training Center (GETC) and Coastal Transit Bill Container (CTBC) algorithm are better than the transmission energy consumed by the maximum transmission power, and the GETC algorithm consumes less energy. When the packet size is 4096 bits, the energy consumed by GETC is 440 MJ, which is only 56% of the energy consumed by the maximum transmitting power. Figure
Figure
Comparison of the average latency of high-priority packets under different loads.
Another simulation scenario is shown in Figure
To compare the delivery rates of high- and low-priority packets, you can force the source node to generate both types of packets at the same time. Figure
Transmission rate of high- and low-priority packets for MAC protocol in emergency and HCL situations.
Figure
Comparison of energy consumption under load over time.
Figure
The sender of a scheduling message receives a conflicting message, depending on the source of its conflicting message, its single-hop neighbour node or its two-hop neighbour node updates the conflicting message and then redistributes the scheduling and broadcasts a new scheduling message to its two-hop inner neighbour node. However, channel conflicts are possible during random storage, and conflicting information may be lost during transmission. If their scheduling messages do not conflict, the neighboring node receiving the scheduling message sends a nonconflicting message to the sender of the scheduling message. To reduce more conflicts, the sender of the scheduling message keeps a list of neighbour nodes; it receives the nonconflicting messages and adds the information from this list to the scheduling message. A neighbour node that does not send a nonconflicting message, if it is in the list, does not receive more nonconflicting messages from its two-hop neighbour node after broadcasting the scheduling message several times, and the sender of the scheduling message conveniently assumes that its scheduling schedule does not conflict with the scheduling schedule of its two-hop neighbour node, and then the node sends the scheduling message directly to the parent node into a notification message (INFORM), and the parent node receives this message and replies Confirmation message, as shown in Figure
WSN control and MAC protocol performance in cyber-physical system.
In microgrid monitoring, the network nodes require high accuracy to collect information, and it is necessary to avoid blind spots due to the death of individual nodes; therefore, it is necessary not only to compare the network survival time under each algorithm, but also to compare the time of the first dead node in the network and the distribution of dead nodes in the network. The relationship between the surviving nodes and the survival period of the three algorithms in Figure
Surviving nodes vs. surviving cycles.
A multihop channel reservation MAC protocol based on a parallel mechanism is proposed, which rationalizes the data transmission and reduces unnecessary idle listening through the channel reservation mechanism. By using the parallel mechanism, the transmission of reservation information and data is carried out simultaneously, saving energy consumption. To a certain extent, it reduces the energy consumption of nodes in WSN, reduces the delay in multihop transmission, and improves the life cycle of the network. The effectiveness of the PCR-MAC protocol is verified by comparing different MAC protocols through the OPNET network simulation tool.
In this paper, based on the analysis of the WSN topology control algorithm, an energy balance topology control algorithm based on a geometric structure is proposed. First of all, we established constraints to ensure network connectivity, under the premise of meeting the constraints, by controlling the transmission power to establish an effective transmission link and reduce unnecessary connections; secondly, based on ensuring network connectivity, the use of planar geometry to draw auxiliary lines, and geometric cutting, the location of the node is divided into different regions, for different regions of the topology of the link. Determine and add some links to enhance the reliability and overall energy optimization of the network; then apply the node energy balancing mechanism to balance the energy consumption of each node and prevent individual nodes from dying due to excessive energy consumption, thus improving the performance of the entire network. A very small controller is designed for the characteristics of the information-physical convergence system. For a large information-physical converged system composed of wireless sensor communication nodes, due to network instability, or suffer from external attacks, it is easy to cause packet loss, out of order, etc., which has a serious impact on the control of physical devices; therefore, this paper defines the attack type as a packet timing attack. Under the interference of this attack, the controller can compensate for the delayed data formation and play a certain inhibitory role in the interference, using the double water tank as the control model, and the simulation results show that it can meet the requirements of stability and robustness of the control. In contrast, the LQG controller clearly cannot play a stable control role in this situation. To improve the energy efficiency and topology robustness of network topology control algorithm. An energy balance topology control algorithm based on geometric cut is proposed. The algorithm firstly performs power control and proposes that the angle between any two neighboring nodes should be less than or equal to 2
The data used to support the findings of this study are available from the corresponding author upon request.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
This research has been financed by the NSFC: research on high efficient video transmission technology based on sparse and low rank decomposition (No. 61671253); general project of Natural Science Research in Universities of Jiangsu Province: research on video transmission technology based on foreground background separation and simultaneous interpreting of information and energy (No. 18kjd510004); Jiangsu Province Education Information Research Project: new interactive education cloud platform design based on the Internet of things, and research on the teaching method of Information turnover in the classroom (No. 20172088); and Jiangsu Province General University Academic Degree Postgraduate scientific research innovation plan project: research on video processing technology based on sparse low rank decomposition and significance detection (No. KYLX16_0661).