Optimization of the energy efficiency in Smart Internet of Vehicles assisted by MEC

Recently, with the evolution from the fifth and beyond fifth generation mobile communication (5G/beyond 5G) [1, 2] to the sixth generation mobile communication (6G)[3], some novel technologies, such as mobile edge computing (MEC) [4, 5], satellite communication [6] and spatial multiplexing [7], have been proposed to promote the development of wireless sensor network (WSN) [8, 9], IoT [10] and other novel applications [11, 12, 13], so as to realize the interconnection of all things. IoV as a special application of IoT attracts extensive attention in terms of road safety, smart transportation and information service [14]. However, the demand of computing resources and battery capacity becomes more and more urgent due to a large number of data traffic of these novel applications in IoV [15]. Hence, mobile edge computing is proposed to mitigate the computing burden of electric vehicles [16]. Meanwhile, the spectrum efficiency is further improved by MIMO and FD technologies [17]. Besides, the energy consumption of data processing will increase significantly due to the large number of data traffic. Therefore, SWIPT is supposed to be a meaningful method which provides sufficient energy for electric vehicles to complete their intensive computing tasks.

The last decade, mobile cloud computing (MCC) [18] as a promising method proposed by researcheres to deal with the limited computing resources of vehicles [19]. However, the network edge meets uncovered service even though massive resources are provided by cloud to afford computing power for vehicles [20]. In addition, long latency resulting from the long propagation between vehicles and remote access unit are significant [21]. In consideration of the problems above, MEC technology is proposed to enhance the computing power at the edge of the network [22, 23, 24]. Hence, the author in [25] investigated the cooperation between MCC and MEC in order to improve the quality of service for vehicles at the network edge. Both [24] and [26] aimed at maximizing the system performance and minimizing the time latency assisted by MEC.

Taking account of the energy limitation of electric vehicles, SWIPT [27] is deployed to extend the battery lifetime of vehicles. Comparing with other practical methods, such as wireless power transfer (WPT) [28] and energy harvesting (EH) [29], SWIPT is used for information decoding and energy harvesting by time-shifting and power splitting, respectively [4]. Besides, MIMO and FD technologies are able to achieve further improvements of spectral efficiency and low latency [30, 31].

Although considerable research has been devoted to improve performance of vehicular network with MEC-assisted technologies, such as a cooperative computing task and redundant data scheme are established to reduce the redundancy of data and allocate the computing tasks through MEC server to abtain the minimal system cost [32]. The author in [33] developed a software defined network enabled heterogeneous vehicular network to improve the scalability of the network as well as provide high reliability and low-latency communication. K. Zhang, et al, [34] proposed a task offloading scheme by jointly consider the data transmission and server selection with Q-learning technology. Based on a multi-objective algorithm, an adaptive strategy of offloading is proposed to optimize the resource allocation in [35]. The author in [36] investigated the secrecy energy efficiency in a distributed massive MIMO Systems. In [37], an alternative optimization problem is also raised to maximize the transmission rate in a 5G-based IoT system by jointly optimizing the node powers and allocation coefficient. Rather less attention has been paid to improve energy efficiency of electric vehicles in IoV considering computing tasks, time latency and spectrum efficiency.

In this paper, we propose an IoV system assisted by MEC which is used for cross-layer offloading to provide low latency and abundant computation resources. Meanwhile, electric vehicles are able to replenish battery capacity by SWIPT technology and the spectrum efficiency is further improved by MIMO and FD technologies. On the premise of computation requirement, time latency and energy constraint of electric vehicles, we maximize the average energy efficiency of electric vehicles. Since the problem is nonconvex, we decouple it into two subproblems at first. According to the previous research [38], a closed-form solution of the central processing unit (CPU) frequency is obtained by solving the first sub-problem. Likewise, in order to tackle with the second sub-problem, three sub-problems are divided from the second sub-problem. Finally, an alternate interior-point iterative scheme is proposed to address these subproblems. Numerical simulations are executed to demonstrate the superior performance of our scheme comparing with the benchmark schemes. Our main contributions in this paper can be summarized as follows:

• •

  We come up an IoV system by using the MEC, the FD and the SWIPT technologies comprehensively, so as to achieve low-latency, abundant computation resources and lower energy consumption by jointly optimizing the CPU frequency, power transfer of electric vehicles, the uplink rate and the offloading tasks.

• •

  Under the same computation task, energy consumption and time latency constraints, the proposed scheme can yield high energy efficiency, lower time transmission and is able to tackle more computing tasks comparing with the benchmark schemes.

• •

  MIMO and FD technologies deployed at anchor node are used to futher improve the spectrum efficiency and the time latency. Meanwhile, we analysis the energy efficiency fairness of the electric vehicles as well.

The remainder of this paper is organized as follows. In section 2, we illustrate the system model. A mathematical problem is formulated and solved in section 3. Section 4 provides numerical results and discussions. Finally, section 5 concludes this paper.

As shown in Fig. 1, we consider an IoV system assisted by MEC, in which a MEC server is deployed at anchor node (AN) and the electric vehicle as a cognitive node connected with AN. Hence, computing tasks will be offloaded partially from electric vehicle nodes (eVNs) to MEC server. Moreover, we assume that the AN works in full-duplex mode which can transmit and receive signals using the same time-frequency resources so as to utilize SWIPT technology to fullfill energy of eVNs as well as receive the offloading computing tasks from the other eVNs at the same time. What’s more, each eVN is equipped with a power splitting receiver for ID and EH with the energy conversion ratio $\beta$ ($\beta\in(0,1)$).

In Fig. 2, each vehicle node $VN_{i}(i\in\{1,...,K\})$ offloads partial computing tasks $m_{i}(0\leq m_{i}\leq M_{i})$ to the AN within the time duration $\tau_{i}^{\text{u}}$, where $M_{i}$ denotes the total computation tasks of $VN_{i}$ and $K$ is the number of vehicle nodes. Simultaneously, AN transmits the computation tasks $m_{j}^{\text{r}}(j\in\{1,...,K\})$ executed by MEC server to $VN_{j}$ within the time duration $\tau_{j}^{\text{d}}$, where $m_{j}^{\text{r}}=\alpha m_{j}(0\leq\alpha\leq 1)$, and $\alpha$ is the ratio of the completed computation tasks from the MEC server. Moreover, the computing time consumption of the MEC server is ignored resulting from the powerful computing ability of it comparing with the vehicle nodes [39]. Furthermore, the local computing tasks $m_{i}^{\text{lo}}$, which equals $M_{i}-m_{i}$, is completed during the time duration $\tau_{i}^{\text{lo}}$. Besides, the uplink time consumption of each VN is assumed to be more than the local computing time consumption of that, which is expressed as $0\leq\tau_{i}^{\text{lo}}\leq\tau_{i}$. More details are given as follows.

Notations: The channel fading is modeled as Rayleigh distribution in this paper. $\bm{h}_{i}^{\text{u}}\in\mathbb{Cc}^{N\times 1}$ and $\bm{h}_{j}^{\text{d}}\in\mathbb{Cc}^{N\times 1}$ satisfy uniform distribution$\sim U(0.5,1)$ which present the uplink and downlink channel, respectively, where $N$ is the antenna numbers of AN, and the superscript u and d refer to uplink and downlink transmission, respectively. Meanwhile, $\bm{H}_{\text{an}}$ represents the self-interference channel of AN. $d_{i}$ and $\bm{d}_{j}\in\mathbb{Cc}^{N\times 1}$ represent the transmitted signal from vehicle node and AN respectively, where $|\bm{d}_{j}|=1$. Moreover, $p_{i}$ and $p_{j}$ denote the transmitting power of VN and AN, respectively. Besides, the additive white Gaussian noise (AWGN) of AN is denoted as $\bm{n}_{\text{a}n}$ with covariance matrix $\sigma_{\text{a}n}^{2}\bm{I}_{M}$. Similarly, $n_{j}$ and $n_{\text{p}s}$ are the AWGN of $VN_{j}$ and PS receiver with variance $\sigma_{j}^{2}$ and $\sigma_{\text{p}s}^{2}$, respectively. Finally, the bandwidth of the system is expressed as $B$.

Since the cost function and the constraints in problem ($P1$) are nonlinear and nonconvex, two subproblems are derived from the problem ($P1$) first. For the first subproblem, a closed-form expression of $f^{\text{t}}$ with fixed variables $\bm{p},\bm{m},\bm{r}$ is obtained. Combining the first subproblem with ($P1$), we derive the second subproblem. For the second subproblem, we utilize an alternate interior-point iterative scheme after decompose it into three subproblems. Finally, the suboptimal varibles $<\bm{f}^{\text{t}},\bm{p},\bm{m},\bm{r}>$ are abtained.