Future Internet, Volume 14, Issue 11 (November 2022) – 44 articles

Integrated Access and Backhaul (IAB) is an enabling technology for efficient 5G millimeter wave (mmWave) New Radio (NR) deployment. The key feature of IAB is multi-hop wireless backhauling, allowing utilizing relaying IAB-nodes to provide cost-efficient access network densification and alleviate the problem of blockages. One of the critical performance measures in such systems is the latency of packet delivery over the multi-hop paths. The paper aims at assessing the impact of multi-hop transmission on the end-to-end delay in an IAB radio access network, taking into account the half-duplex constraint. We build a detailed queuing theory model for latency assessment in time-division-multiplexing (TDM)-based IAB deployments and evaluate the delay due to queuing in the network nodes for several cell topologies and under different time allocation strategies between access and backhaul. The paper considers a practical Manhattan-style urban deployment, which is characteristically impaired by the blockage of buildings. The numerical results show that balancing the access and backhaul micro phases is crucial for reducing the end-to-end packet delay, at least in the uplink, while increasing the number of network hops yields a linear increase in the total packet delay for both the uplink and downlink. The numerical results were obtained via simulation using the open-source software OMNeT++. Full article

Data reconciliation in general, and filesystem synchronization in particular, lacks rigorous theoretical foundation. This paper presents, for the first time, a complete analysis of synchronization for two replicas of a theoretical filesystem. Synchronization has two main stages: identifying the conflicts, and resolving them. All existing (both theoretical and practical) synchronizers are operation-based: they define, using some rationale or heuristics, how conflicts are to be resolved without considering the effect of the resolution on subsequent conflicts. Instead, our approach is declaration-based: we define what constitutes the resolution of all conflicts, and for each possible scenario we prove the existence of sequences of operations/commands which convert the replicas into a common synchronized state. These sequences consist of operations rolling back some local changes, followed by operations performed on the other replica. The set of rolled-back operations provides the user with clear and intuitive information on the proposed changes, so she can easily decide whether to accept them or ask for other alternatives. All possible synchronized states are described by specifying a set of conflicts, a partial order on the conflicts describing the order in which they need to be resolved, as well as the effect of each decision on subsequent conflicts. Using this classification, the outcomes of different conflict resolution policies can be investigated easily. Full article

Interactivity has been a very sought-after feature in professional journalism ever since the media industry transitioned from print into the online space. Within this context, chatbots started to infiltrate the media sphere and provide news organizations with new and innovative ways to create and share their content, with an even larger emphasis on back-and-forth communication and news reporting personalization. The present research highlights two important factors that can determine the efficient integration of chatbots in professional journalism: the feasibility of chatbot programming by journalists without a background in computer science using coding-free platforms and the usability of the created chatbot agents for news reporting to the audience. This paper aims to review some of the most popular, coding-free chatbot creation platforms that are available to journalists today. To that end, a three-phase evaluation framework is introduced. First off, the interactivity features that they offer to media industry workers are evaluated using an appropriate metrics framework. Secondly, a two- part workshop is conducted where journalists use the aforementioned platforms to create their own chatbot news reporting agents with minimum training, and lastly, the created chatbots are evaluated by a larger audience concerning the usability and overall user experience. Full article

The ongoing deployment of 5G networks is seen as a key enabler for realizing upcoming interconnected services at scale, including the massive deployment of the Internet of Things, providing V2X communications to support autonomous vehicles, and the increase in smart homes, smart cities, and Industry 4 [...] Full article

The real-world use cases of blockchain technology, such as faster cross-border payments, identity management, smart contracts, cryptocurrencies, and supply chain–blockchain technology are here to stay and have become the next innovation, just like the Internet. There have been attempts to formulate digital money, but they have not been successful due to security and trust issues. However, blockchain needs no central authority, and its operations are controlled by the people who use it. Furthermore, it cannot be altered or forged, resulting in massive market hype and demand. Blockchain has moved past cryptocurrency and discovered implementations in other real-life applications; this is where we can expect blockchain technology to be simplified and not remain a complex concept. Blockchain technology’s desirable characteristics are decentralization, integrity, immutability, verification, fault tolerance, anonymity, audibility, and transparency. We first conduct a thorough analysis of blockchain technology in this paper, paying particular attention to its evolution, applications and benefits, the specifics of cryptography in terms of public key cryptography, and the challenges of blockchain in distributed transaction ledgers, as well as the extensive list of blockchain applications in the financial transaction system. This paper presents a detailed review of blockchain technology, the critical challenges faced, and its applications in different fields. Blockchain in the transaction system is explained in detail with a summary of different cryptocurrencies. Some of the suggested solutions are given in the overall study of the paper. Full article

In today’s world, phishing attacks are gradually increasing, resulting in individuals losing valuables, assets, personal information, etc., to unauthorized parties. In phishing, attackers craft malicious websites disguised as well-known, legitimate sites and send them to individuals to steal personal information and other related private details. Therefore, an efficient and accurate method is required to determine whether a website is malicious. Numerous methods have been proposed for detecting malicious uniform resource locators (URLs) using deep learning, machine learning, and other approaches. In this study, we have used malicious and benign URLs datasets and have proposed a detection mechanism for detecting malicious URLs using recurrent neural network models such as long short-term memory (LSTM), bidirectional long short-term memory (Bi-LSTM), and the gated recurrent unit (GRU). Experimental results have shown that the proposed mechanism achieved an accuracy of 97.0% for LSTM, 99.0% for Bi-LSTM, and 97.5% for GRU, respectively. Full article

Identity federation allows one to link a user’s digital identities across several identity management systems. Federated identity management (FIM) ensures that users have easy access to the available resources. However, scaling FIM to numerous partners is a challenging process due to the interoperability issue between different federation architectures. This study proposes a dynamic identity federation model to eliminate the manual configuration steps needed to establish an organizational identity federation by utilizing the OpenID Connect (OIDC) framework. The proposed model consists of three major steps to establish dynamic FIM: first, the discovery of the OpenID service provider, which indicates the location of the partner organization; second, the registration of the OpenID relying party, which allows the organization and its partner to negotiate information for establishing the federation; finally, establishing the dynamic trust federation. The proposed dynamic FIM model allows applications to provide services to end-users coming from various domains while maintaining a trust between clients and service providers. Through our proposed dynamic identity federation model, organizations can save hundreds of hours by achieving dynamic federation in runtime and serving a large number of end-users. Full article

Dynamic and smart Internet of Things (IoT) infrastructures allow the development of smart healthcare systems, which are equipped with mobile health and embedded healthcare sensors to enable a broad range of healthcare applications. These IoT applications provide access to the clients’ health information. However, the rapid increase in the number of mobile devices and social networks has generated concerns regarding the secure sharing of a client’s location. In this regard, federated learning (FL) is an emerging paradigm of decentralized machine learning that guarantees the training of a shared global model without compromising the data privacy of the client. To this end, we propose a K-anonymity-based secure hierarchical federated learning (SHFL) framework for smart healthcare systems. In the proposed hierarchical FL approach, a centralized server communicates hierarchically with multiple directly and indirectly connected devices. In particular, the proposed SHFL formulates the hierarchical clusters of location-based services to achieve distributed FL. In addition, the proposed SHFL utilizes the K-anonymity method to hide the location of the cluster devices. Finally, we evaluated the performance of the proposed SHFL by configuring different hierarchical networks with multiple model architectures and datasets. The experiments validated that the proposed SHFL provides adequate generalization to enable network scalability of accurate healthcare systems without compromising the data and location privacy. Full article

Modern Public Key Infrastructures (PKIs) allow users to create and maintain centrally stored cryptographic certificates. These infrastructures use a so-called certificate chain. At the root of the chain, a root Certification Authority (CA) is responsible for issuing the base certificate. Every verification and certification step within the chain is based upon the security of said root CA. Thus, its operation security is of great concern. Since the root certificates are stored locally on the root CA, any Denial of Service (DoS) attack may render the whole certificate chain, which is based on of the attacked root CA, inoperable. Therefore, this article evaluates different approaches to a decentralized data storage system that is based on the Distributed Ledger Technology (DLT). To show the real-world potential of the proposed approaches, we also evaluate the different technologies using a novel PKI mechanism called Near Field Communication Key Exchange (NFC-KE). The results indicate that modern distributed data storage solutions such as Interplanetary Filesystem (IPFS) and SIA can have significant performance and decentralization benefits in comparison to purely Blockchain-based technologies like Hyperledger Fabric. However, they lack any Smart Contract functionality, which requires a software developer to implement verification mechanisms in centralized software solutions. Full article

The first week of August 2022 saw the world’s cryptographers grapple with the second shocker of the year. Another one of the four post-quantum cryptography (PQC) algorithms selected by the NIST (National Institute of Standards and Technology) in a rigorous 5-year process was cracked by a team from Belgium. They took just 62 min and a standard laptop to break the PQC algorithm to win a USD 50,000 bounty from Microsoft. The first shocker came 6 months earlier, when another of the NIST finalists (Rainbow) was taken down. Unfortunately, both failed PQC algorithms are commercially available to consumers. With 80 of the 82 PQC candidates failing the NIST standardization process, the future of the remaining two PQC algorithms is, at best, questionable, placing the rigorous 5-year NIST exercise to build a quantum-safe encryption standard in jeopardy. Meanwhile, there is no respite from the quantum threat that looms large. It is time we take a step back and review the etiology of the problem de novo. Although state-of-the-art computer security heavily relies on cryptography, it can indeed transcend beyond encryption. This paper analyzes an encryption-agnostic approach that can potentially render computers quantum-resistant. Zero-vulnerability computing (ZVC) secures computers by banning all third-party permissions, a root cause of most vulnerabilities. ZVC eliminates the complexities of the multi-layered architecture of legacy computers and builds a minimalist, compact solid-state software on a chip (3SoC) that is robust, energy-efficient, and potentially resistant to malware as well as quantum threats. Full article

Cloud–fog computing is a large-scale service environment developed to deliver fast, scalable services to clients. The fog nodes of such environments are distributed in diverse places and operate independently by deciding on which data to process locally and which data to send remotely to the cloud for further analysis, in which a Service-Level Agreement (SLA) is employed to govern Quality of Service (QoS) requirements of the cloud provider to such nodes. The provider experiences varying incoming workloads that come from heterogeneous fog and Internet of Things (IoT) devices, each of which submits jobs that entail various service characteristics and QoS requirements. To execute fog workloads and meet their SLA obligations, the provider allocates appropriate resources and utilizes load scheduling strategies that effectively manage the executions of fog jobs on cloud resources. Failing to fulfill such demands causes extra network bottlenecks, service delays, and energy constraints that are difficult to maintain at run-time. This paper proposes a joint energy- and QoS-optimized performance framework that tolerates delay and energy risks on the cost performance of the cloud provider. The framework employs scheduling mechanisms that consider the SLA penalty and energy impacts of data communication, service, and waiting performance metrics on cost reduction. The findings prove the framework’s effectiveness in mitigating energy consumption due to QoS penalties and therefore reducing the gross scheduling cost. Full article

Sentiment analysis of online Chinese buzzwords (OCBs) is important for healthy development of platforms, such as games and social networking, which can avoid transmission of negative emotions through prediction of users’ sentiment tendencies. Buzzwords have the characteristics of varying text length, irregular wording, ignoring syntactic and grammatical requirements, no complete semantic structure, and no obvious sentiment features. This results in interference and challenges to the sentiment analysis of such texts. Sentiment analysis also requires capturing effective sentiment features from deeper contextual information. To solve the above problems, we propose a deep learning model combining BERT and BiLSTM. The goal is to generate dynamic representations of OCB vectors in downstream tasks by fine-tuning the BERT model and to capture the rich information of the text at the embedding layer to solve the problem of static representations of word vectors. The generated word vectors are then transferred to the BiLSTM model for feature extraction to obtain the local and global semantic features of the text while highlighting the text sentiment polarity for sentiment classification. The experimental results show that the model works well in terms of the comprehensive evaluation index F1. Our model also has important significance and research value for sentiment analysis of irregular texts, such as OCBs. Full article

It is meaningful to analyze the market correlations for stock selection in the field of financial investment. Since it is difficult for existing deep clustering methods to mine the complex and nonlinear features contained in financial time series, in order to deeply mine the features of financial time series and achieve clustering, a new end-to-end deep clustering method for financial time series is proposed. It contains two modules: an autoencoder feature extraction network based on TCN (temporal convolutional neural) networks and a temporal clustering optimization algorithm with a KL (Kullback–Leibler) divergence. The features of financial time series are represented by the causal convolution and the dilated convolution of TCN networks. Then, the pre-training results based on the KL divergence are fine-tuned to make the clustering results discriminative. The experimental results show that the proposed method outperforms existing deep clustering and general clustering algorithms in the CSI 300 and S&P 500 index markets. In addition, the clustering results combined with an inference strategy can be used to select stocks that perform well or poorly, thus guiding actual stock market trades. Full article

Many IoT-blockchain systems in which blockchain connections run on an infrastructure-based network, such as Wi-Fi or LTE, face a severe problem: the single point of failure (SPoF) (i.e., depending on the availability, an access point of an LTE base station). Using infrastructure-less networks (i.e., ad hoc networks) is an efficient approach to prevent such highly disruptive events. An ad hoc network can automatically restore blockchain communication using an ad hoc routing protocol, even if a node fails. Moreover, an ad hoc routing protocol is more efficient when considering the IoT nodes’ mobility. In this paper, we first construct IoT-blockchain systems on emulated and real ad hoc networks with Ethereum and three ad hoc routing protocols (i.e., OLSR, BATMAN, and BABEL). We then evaluate the blockchain recovery time in static and mobile scenarios. The results show that BATMAN achieves the best blockchain recovery performance in all investigated scenarios because BATMAN only determines whether to switch a route by comparing the number of OGM packets received from a different next-hop. More specifically, in the small-scale real IoT-blockchain, BATMAN recovers at least 73.9% and 59.8% better than OLSR and BABEL, respectively. In the medium-scale emulated IoT-blockchain, the recovery time of BATMAN is at least 69% and 60% shorter than OLSR or BABEL, respectively. Full article

The growth and impact of video games in education at an international level is a reality. Research shows that gamers can increase their knowledge, skills, and behavioural flexibility. However, there has been no in-depth research into the relationship between current video games and the key competences for lifelong learning set out by the European Commission. This research focuses on learning acquisition through playing the popular game “Animal Crossing: New Horizons”. The Communicative Methodology has been used in this research through, on the one hand, use of the Social Impact in Social Media (SISM) method involving the analysis of 1000 comments posted on the social network Twitter and, on the other hand, through communicative inter, sanviews with five gamers and a family member of a user. The results show that the Animal Crossing video game promotes learning achievements regarding literacy, multilingualism, mathematical skills, digital competence, social skills, citizenship, entrepreneurship, and cultural awareness. Full article

Even though both internet-of-cloud services and biometric systems (BSs) are used worldwide, popular and expanded among public and private services, their utilization has not been extended among Greek users, and in particular within Greek HEIs. This paper aims to examine the University of the Aegean (UA) students’ perceptions, especially on BS acceptance, determining the impact of their sociodemographic characteristics toward this. An online instrument was developed based on constructs of technology acceptance models, and previous research focused on BS acceptance, including a great variety of sociodemographic variables. The authors collected 768 complete responses to explore the social and demographic attributes that impact on students’ BS acceptance through exploratory descriptive research. This study found that several UA students’ sociodemographic attributes are linked to their BS acceptance. Specifically, gender, age, study level, year of study, professional profile and fathers’ occupation are useful to determine which students’ groups are in favor of BS, while employment status, annual income and mothers’ occupation are weak toward this. Security and privacy concerns are indicated to influence users’ BS acceptance as well. This study provides a unique approach to better understanding the Greek students’ BS acceptance, with important implications for improved BS marketing in Greece in synergy with the internet-of-cloud services and their potential adoption in HEIs internal settings. It also provides developers with further knowledge for security and privacy non-functional requirements in order to increase users’ acceptance and to address these challenges within the internet of cloud. Full article

Technological evolution and its potential impacts are of significant interest to governments, corporate organizations and for academic enquiry; but assessments of technology progression are often highly subjective. This paper prototypes potential objective measures to assess technology progression using internet-based data. These measures may help reduce the subjective nature of such assessments and, in conjunction with other techniques, reduce the uncertainty of technology progression assessment. The paper examines one part of the technology ecosystem, namely, academic research and publications. It uses analytics performed against a large body of academic paper abstracts and metadata published over 20 years to propose and demonstrate candidate indicators of technology progression. Measures prototyped are: (i) overall occurrence of technologies used over time in research, (ii) the fields in which this use was made; (iii) the geographic spread of specific technologies within research and (iv) the clustering of technology research over time. An outcome of the analysis is an ability to assess the measures of technology progression against a set of inputs and a set of commentaries and forecasts made publicly in the subject area over the last 20 years. The potential automated indicators of research are discussed together with other indicators which might help working groups in assessing technology progression using more quantitative methods. Full article

With the blooming of blockchain-based smart contracts in decentralized applications, the security problem of smart contracts has become a critical issue, as vulnerable contracts have resulted in severe financial losses. Existing research works have explored vulnerability detection methods based on fuzzing, symbolic execution, formal verification, and static analysis. In this paper, we propose two static analysis approaches called ASGVulDetector and BASGVulDetector for detecting vulnerabilities in Ethereum smart contacts from source-code and bytecode perspectives, respectively. First, we design a novel intermediate representation called abstract semantic graph (ASG) to capture both syntactic and semantic features from the program. ASG is based on syntax information but enriched by code structures, such as control flow and data flow. Then, we apply two different training models, i.e., graph neural network (GNN) and graph matching network (GMN), to learn the embedding of ASG and measure the similarity of the contract pairs. In this way, vulnerable smart contracts can be identified by calculating the similarity to labeled ones. We conduct extensive experiments to evaluate the superiority of our approaches to state-of-the-art competitors. Specifically, ASGVulDetector improves the best of three source-code-only static analysis tools (i.e., SmartCheck, Slither, and DR-GCN) regarding the F1 score by 12.6% on average, while BASGVulDetector improves that of the three detection tools supporting bytecode (i.e., ContractFuzzer, Oyente, and Securify) regarding the F1 score by 25.6% on average. We also investigate the effectiveness and advantages of the GMN model for detecting vulnerabilities in smart contracts. Full article

The development of ubiquitous computing technology and the emergence of XR could provide pedestrian navigation with more options for user interfaces and interactions. In this work, we aim investigate the role of a mixed-reality map interface in urban exploration to enhance pedestrians’ mental satisfaction. We propose a mixed-reality 3D minimap as a part of the navigation interface which pedestrians could refer to and interact during urban exploration. To further explore the different levels of detail of the map interface, we conducted a user study (n = 28, two groups with two tasks). We designed two exploratory activities as experimental tasks with two map modes (a normal one and a simplified one) to discuss the detailed design of the minimap interface. The results indicated that participants showed a positive attitude toward our method. The simplified map mode could result in a lower perceived workload in both tasks while enhancing performance in specific navigation, such as wayfinding. However, we also found that pedestrians’ preference for the level of detail of the minimap interface is dynamic in navigation. Thus, we suggest discussing the different levels of detail further in specific scenarios. Finally, we also summarize some findings observed during user study for inspiring the study of virtual map interface of future mixed-reality navigation for urban exploration in various scenarios. Full article

The increasing prevalence of marine pollution during the past few decades motivated recent research to help ease the situation. Typical water quality assessment requires continuous monitoring of water and sediments at remote locations with labour-intensive laboratory tests to determine the degree of pollution. We propose an automated water quality assessment framework where we formalise a predictive model using machine learning to infer the water quality and level of pollution using collected water and sediments samples. Firstly, due to the sparsity of sample collection locations, the amount of sediment samples of water is limited, and the dataset is incomplete. Therefore, after an extensive investigation on various data imputation methods’ performance in water and sediment datasets with different missing data rates, we chose the best imputation method to process the missing data. Afterwards, the water sediment sample will be tagged as one of four levels of pollution based on some guidelines and then the machine learning model will use a specific technique named classification to find the relationship between the data and the final result. After that, the result of prediction can be compared to the real result so that it can be checked whether the model is good and whether the prediction is accurate. Finally, the research gave improvement advice based on the result obtained from the model building part. Empirically, we show that our best model archives an accuracy of 75% after accounting for 57% of missing data. Experimentally, we show that our model would assist in automatically assessing water quality screening based on possibly incomplete real-world data. Full article

Within the scope of precision agriculture, many applications have been developed to support decision making and yield enhancement. Fruit detection has attracted considerable attention from researchers, and it can be used offline. In contrast, some applications, such as robot vision in orchards, require computer vision models to run on edge devices while performing inferences at high speed. In this area, most modern applications use an integrated graphics processing unit (GPU). In this work, we propose the use of a tensor processing unit (TPU) accelerator with a Raspberry Pi target device and the state-of-the-art, lightweight, and hardware-aware MobileDet detector model. Our contribution is the extension of the possibilities of using accelerators (the TPU) for edge devices in precision agriculture. The proposed method was evaluated using a novel dataset of peaches with three cultivars, which will be made available for further studies. The model achieved an average precision (AP) of 88.2% and a performance of 19.84 frames per second (FPS) at an image size of 640 × 480. The results obtained show that the TPU accelerator can be an excellent alternative for processing on the edge in precision agriculture. Full article

Modern internet of things (IoT) applications can benefit from advanced communication paradigms, including multicast and anycast. Next-generation internet architectures, such as information-centric networking (ICN), promise to support these paradigms, but at the same time they introduce new security challenges. This paper presents a solution that extends an ICN-like architecture based on software-defined networking (SDN) that supports those communication paradigms. Using the proposed solution, the underlying architecture is enhanced with a novel security mechanism that allows content “advertisements” only from authorized endpoints. This mechanism prevents “content pollution”, which is a significant security threat in ICN architectures. The proposed solution is lightweight, and it enables identity sharing as well as secured and controlled identity delegation. Full article

The blockchain is an emerging technology that has the potential to revolutionize the gaming industry among a wide range of different business fields. So far, only a few studies have been conducted about blockchain gaming. This study introduces a mobile game utilizing blockchain asset tokens and smart contracts. It was developed for research purposes and used to demonstrate blockchain-based games using semi-structured interviews. This study follows the exploratory research paradigm, which aims to map research of little-known areas. This study focuses on how participants perceived blockchain attributes such as trust, transparency, and user-generated content and how this affected engagement and their willingness to play the game again. Based on our evaluation, generating blockchain assets positively impacted player retention. According to the results, providing genuine asset ownership through the blockchain contributes to environmental engagement and self-engagement, as well as player retention. Another positive blockchain feature discovered from the interview data is user-generated content implementation into games. Full article

Intentionally deceptive online content represents a significant issue for society. Multiple techniques have been proposed to identify and combat its spread. To understand how to inform individuals most effectively about the potential biases of and other issues with content, this article studies factors that impact the perception of online content. Specifically, it looks at how these factors have similar or different impact depending on the income level, political affiliation and online usage time of Americans. A national survey was conducted that asked respondents about multiple factors that influence their and others’ perception of online content trustworthiness. It also asked what the ideal impact of these factors should be. This data is presented and analyzed herein, conclusions are drawn and their implications, with regard to preventing the spread of deceptive online content, are discussed. Full article

The violation traffic laws by driving at high speeds, the overloading of passengers, and the unfastening of seatbelts are of high risk and can be fatal in the event of any accident. Several systems have been proposed to improve passenger safety, and the systems either use the sensor-based approach or the computer-vision-based approach. However, the accuracy of these systems still needs enhancement because the entire road network is not covered; the approaches utilize complex estimation techniques, and they are significantly influenced by the surrounding environment, such as the weather and physical obstacles. Therefore, this paper proposes a novel IoT-based traffic violation monitoring system that accurately estimates the vehicle speed, counts the number of passengers, and detects the seatbelt status on the entire road network. The system also utilizes edge computing, fog computing, and cloud computing technologies to achieve high accuracy. The system is evaluated using real-life experiments and compared with another system where the edge and cloud layers are used without the fog layer. The results show that adding a fog layer improves the monitoring accuracy as the accuracy of passenger counting rises from 94% to 97%, the accuracy of seatbelt detection rises from 95% to 99%, and the root mean square error of speed estimation is reduced from 2.64 to 1.87. Full article

With the wider adoption of edge computing services, intelligent edge devices, and high-speed V2X communication, compute-intensive tasks for autonomous vehicles, such as object detection using camera, LiDAR, and/or radar data, can be partially offloaded to road-side edge servers. However, data privacy becomes a major concern for vehicular edge computing, as sensitive sensor data from vehicles can be observed and used by edge servers. We aim to address the privacy problem by protecting both vehicles’ sensor data and the detection results. In this paper, we present vehicle–edge cooperative deep-learning networks with privacy protection for object-detection tasks, named vePOD for short. In vePOD, we leverage the additive secret sharing theory to develop secure functions for every layer in an object-detection convolutional neural network (CNN). A vehicle’s sensor data is split and encrypted into multiple secret shares, each of which is processed on an edge server by going through the secure layers of a detection network. The detection results can only be obtained by combining the partial results from the participating edge servers. We have developed proof-of-concept detection networks with secure layers: vePOD Faster R-CNN (two-stage detection) and vePOD YOLO (single-stage detection). Experimental results on public datasets show that vePOD does not degrade the accuracy of object detection and, most importantly, it protects data privacy for vehicles. The execution of a vePOD object-detection network with secure layers is orders of magnitude faster than the existing approaches for data privacy. To the best of our knowledge, this is the first work that targets privacy protection in object-detection tasks with vehicle–edge cooperative computing. Full article