In view of the new service scenarios and the demand for high intelligence network in computing and network convergence (CNC), the concept of autonomous CNC (Auto-CNC) is elaborated, where intelligence was introduced into all the aspects of CNC, including resource integration, process automation, and system intelligence.The current research directions and remaining challenges of CNC were introduced, and three key features, i.e., intent-driven computing network, the autonomous system operation and the adaptive co-evolution of communication, computing intelligence, were summarized from the proposed Auto-CNC.Meanwhile, the reference architecture and key technologies of Auto-CNC were described, which were followed by several preliminary exploration cases.Finally, future research trends and technical advice were discussed and recommended.

With the development of the internet of things, a large number of user device (UD) were connected to cellular network.Since the changes in the spatial distribution of UD and application requirements, it is necessary to dynamically adjust the UD’ offloading decision.Comprehensively considering various parameter information in the networks such as the spatial distribution of UD, application requirements, and the processing capability of the edge servers on the base station (BS) side, the offloading decision of UD were optimized from the perspective of distribution.Based on the optimal transport theory, a delay optimization algorithm was proposed to reduce the average delay of the UD’ computing tasks offloading process by reasonably planning the offloading BS of the UD in the networks.The simulation results show that the average delay can be reduced by 81.06% using the proposed offloading mechanism based on delay optimization, and the traffic handled by each BS is balanced.

A low complexity downlink channel estimation method for high-speed mobile scenarios was proposed for the extra-large scale multiple-input multiple-output (MIMO) orthogonal time-frequency space (OTFS) system.Different from the existing studies, the proposed method considers the significant spatial non-stationary characteristics of extra-large scale MIMO-OTFS system.Based on the visible path region, an enhanced sparse orthogonal matching pursuit (OMP) algorithm with low complexity was proposed, uplink-assisted downlink channel estimation was achieved using the mapping relationship between the uplink and downlink channels in the frequency-division duplex (FDD) mode.Simulation results show that the proposed uplink-assisted downlink channel estimation method can accurately represent the non-stationary of the system, and achieve significant improvement in the channel estimation performance while reducing the computational complexity , and still perform well in high-mobility IoT scenarios.

For a multi-user integrated sensing and communication system in the network of vehicles, a robust sensing-assisted communication massive multiple-input multiple-output (MIMO) orthogonal time frequency space (OTFS) transmission scheme was proposed.Due to the limited sensing accuracy of the radar, errors existed in the channel state information (CSI) reconstructed based on sensing parameters.The transmission performance decreased as a result.To address this issue, the CSI in the delay doppler domain was reconstructed based on the sensing parameters by the transmitter firstly.And a robust beam forming scheme was designed considering the CSI error.Secondly, the channel estimation error and inter user interference were perceived by receivers based on sensing parameters.Then the robust receiver was designed by incorporating the perceived interference errors into the signal detector in an analytical way.Finally, numerical simulation results show that the proposed method effectively reduces the system bit error rate and increases the data reception rate of users.The proposed method improves the overall system performance in this situation.

Time-sensitive networking (TSN) is a set of protocols developed by the IEEE TSN task group, aiming at achieving deterministic communications over Ethernet.As the implementation method of TSN traffic scheduling is not specified in the protocols, the routing and scheduling algorithm for TSN remains an open issue.The joint optimization problem of routing and scheduling in TSN for industrial applications was modeled, and then an online heuristic algorithm was proposed to deliver the routing and scheduling solution for dynamic traffics.The routing path was determined by optimizing both the transmission delay and network load factors, and the scheduling time was quickly conducted by twice clipping operations.Finally, a simulated TSN testbed was developed with NeSTiNg framework based on OMNeT.The simulation results show that the execution time of the proposed algorithm outperforms the baseline algorithms even with large scale of network size and network traffics.It shows that the proposed algorithm guarantees the real-time performance even in dynamically changing networks.

Ultrawideband (UWB) is a hot technology for indoor positioning with large bandwidth, strong anti-interference ability, and high multipath resolution capacity.However, due to the complex indoor environment, UWB signal propagation will inevitably be blocked, resulting in non-line-of-sight (NLOS) propagation, which greatly reduces the accuracy of UWB positioning.Therefore, identifying NLOS signals accurately and discarding or correcting them are important to alleviate the problem of the decline in positioning accuracy.The majority of present NLOS identification work focuses on scenes with building structures such as walls.Further discussion is needed for scenes obscured by pedestrians.Since the impact of human obstacles on the signals is more complex and cannot be ignored, the NLOS identification under pedestrian occlusion was studied.By comparing a variety of machine learning methods and signal feature combinations, the random forest method based on the three-dimensional features of the first path signal power, the received signal power, and the measured distance was proposed.These features with fewer dimensions and easy extraction were used to achieve a high identification percentage for NLOS.The experimental results based on the measured data of different devices show that the NLOS identification accuracy based on the proposed method reaches 99.05%, 99.32% and 98.81% respectively.

Aiming at large-scale time-sensitive networking (TSN) scenarios that can only ensure frequency synchronization between devices, a hybrid traffic scheduling mechanism was proposed based on time-aware shaper (TAS) combined with cycle specified queuing and forwarding (CSQF).Firstly, a scheme was investigated for achieving periodic cyclic mapping alignment between two adjacent nodes located in different time domains.Secondly, combining segment routing technology, a heuristic algorithm based on CSQF mechanism for joint routing and scheduling was proposed to complete resource allocation for large bandwidth traffic.The experimental results show that the proposed scheme significantly improves the system resource utilization and scheduling success rate compared to existing research results, achieving efficient TSN traffic scheduling in wide area network scenarios.

The low security and poor privacy of the data in edge computing restrict the development of edge computing.Block chains can provide security for data in edge computing using their own tamper resistance, while protecting privacy by use of traceability.But the bottleneck of blockchain's scalability has become a barrier to their application in the field of edge computing.To solve the problem that blockchain can not meet the needs of a large number of nodes to process data at the same time when applied to edge computing, a two-layer sharding scheme was presented, which meets the needs of edge computing scenarios.Geographic location-based partitioning of nodes was implemented using the improved K-means algorithm, and a local blockchain network consensus (LBNC) algorithm was designed based on the idea of delegated proof of stake (DPoS) and practical Byzantine fault tolerance (PBFT).Simulation results show that the proposed scheme has less delay and higher throughput than those of PBFT, and the total throughput increases with the number of shards.

A ZigBee network anonymous authentication scheme based on Chebyshev chaotic mapping and Chinese remainder theorem (CRT) was proposed to solve the problems such as the incomplete reliability of ZigBee network trust center and the lack of identity authentication when accessing the network.The proposed scheme can not only realize two-way authentication of anonymous identity, but also ensure the security of key distribution when ZigBee network structure changes dynamically.It is mainly based on a ZigBee and NB-IoT wireless heterogeneous gateway, so that the server can effectively manage the nodes in the network through this gateway.From security analysis and comparison with other related literature, the proposed scheme has higher security, with anonymity and unlink ability.In addition, the results show that the proposed scheme has more advantages than other schemes on the computational overhead.

The wireless power transfer technique is promising in solving the energy bottleneck of sensor nodes in wireless sensor networks, which can thus prolong the network lifetime or even maintain sustainable network operations.Most existing works focused on optimizing the static chargers’ deployment or mobile chargers’ charging path for static sensor nodes with fixed sensor node positions, ignoring the scenario with mobile sensor nodes.Thus, design and optimize the charging path of a mobile charger was studied for dynamic wireless sensor networks with mobile sensor nodes, to maximize the charging utility within a finite time horizon, that is, the charger can encounter as more sensor nodes as possible in a limited time and charge them.Notice that the mobile charger may stop to simultaneously charge multiple nodes within its charging range during its charging tour.The proposed charging path optimization problem was proven to be an APX-hard problem.Then, based on the constructed directed acyclic graph using discretization method, a layer-wise pruning algorithm based on the backtracking method was proposed.The proposed algorithm took the solution generated by the greedy algorithm as the benchmark and searched the optimal charging path under a fixed time division by layer-wise pruning.Simulation results show that the proposed algorithm can effectively improve the charging utility .

The unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) network, where the UAV can be used as a mobile base station to collect data from the multiple user devices was studied.Since it should satisfy the UAV velocity constraint, the system computation latency constraint, and the system communication latency constraint, a joint scheme of bandwidth allocation and UAV 3D trajectory was proposed to minimize the total energy consumption of the network.The corresponding non-convex optimization problem that is difficult to solve was converted into two sub-optimization problems, which are the UAV 3D trajectory optimization subproblem by fixing bandwidth allocation, and the user equipment bandwidth allocation subproblem by fixing UAV 3D trajectory.Experiments demonstrate that the energy consumption of the proposed algorithm is less than that of the other two typical algorithms.Furthermore, the theoretical basis for solving the limited resource and energy problem in UAV-assisted MEC network was provided.

In modern agriculture IoT systems, edge computing is an indispensable component.In this context, it is feasible to deploy lightweight pest and disease image recognition tasks on edge devices.However, due to the constraints of device computation and storage capabilities, this task faces significant challenges.To address these challenges, an economically practical method was proposed for pest and disease image recognition on STM32 edge devices.Specifically, the MobileNetv2 structure was improved to better suit the characteristics of STM32, quantization-aware training technique was used to compresses the network, model portability was enhanced.Meanwhile, the X-CUBE-AI was used to arrange the model and evaluate the performance.Experimental results demonstrate that the proposed model not only ensures image classification accuracy but also reduces the Flash and RAM resource consumption on STM32 compared to other lightweight networks.

To address the problem of low detection accuracy caused by the difficulty in effectively capturing significant and diversified feature information for packaging deformation and overlap products, a location learnable visual center (LLVC) mechanism was designed to improve YOLOX-s, achieving higher detection accuracy.To effectively deal with product packaging deformation and overlap phenomena, firstly, global context information was captured through a lightweight multi-layer perceptron to help the model better understand spatial information in product features.Secondly, the local feature representation ability was enhanced by the designed LLVC and the spatial information was used to allocate learnable weights for local features to increase the attention of discriminative local features.Finally, the intersection over union (IoU) loss function was replaced with centered intersection over union (CIoU) and power parameters were introduced on this basis to effectively reduce the missed detection rate.Experimental results show that the proposed method achieves an accuracy of 91.3% on the retail product checkout (RPC) dataset, which is 2.2% higher than YOLOX-s and better than current mainstream lightweight object detection algorithms.At the same time, frame per second (FPS) is 97 frame/s, and the model size is 9.48 MB.It can accurately and in real-time detect retail products in scenarios where computing resources are limited.

Human activity sensing based on Wi-Fi channel state information (CSI) has an important application prospect in future intelligent interaction scenarios such as virtual reality, intelligent games, and the metaverse.Accurate sensing of complex and continuous human activities is an important challenge for Wi-Fi sensing.Convolutional neural network (CNN) has the ability of spatial feature extraction but is poor at modeling the temporal features of the data.While long short-term memory (LSTM) network or gated recurrent unit (GRU) network, which are suitable for modeling time-series data, neglect learning spatial features of data.In order to solve this problem, an improved CNN that integrates bidirectional gated recurrent unit (BiGRU) network was proposed.The bi-directional feature extraction ability of BiGRU was used to capture the correlation and dependence of the front and back information in the time series data.The extraction of the spatiotemporal features of the time series CSI data was realized, and then the mapping relationship between the action and the CSI data was present.Thus the recognition accuracy of the complex continuous action was improved.The proposed network structure was tested with basketball actions.The results show that the recognition accuracy of this method is above 95% under various conditions.Compared with the traditional multi-layer perceptron (MLP), CNN, LSTM, GRU, and attention based bidirectional long short-term memory (ABLSTM) baseline methods, the recognition accuracy has been improved by 1%～20%.

The internet of things (IoT) has gained wide adoption across various industries, driving digitalization and intelligence in industry applications.However, the data collected by IoT devices in some industry applications may be closely linked to user privacy and property security.To ensure the security of such data, a cost-effective, multifunctional hardware architecture design based on the Chinese authenticated encryption algorithm SM4-GCM (Galois/Counter Mode) was proposed, which offered a balanced approach to performance, cost, and hardware-level data confidentiality and integrity assurance, and supported three operation modes: SM4-CTR, SM4-ECB, and SM4-GCM.The implementation results on the field programmable gate array (FPGA) development board demonstrate that the design requires only 1 761 look-up tables and 1 825 registers, occupies only 604 slices, and has a throughput rate of 39.78 Mbit/s@100 MHz.These results suggest that the proposed design can effectively meet the requirements of IoT data security applications.