Computation offloading algorithm for cloud robot based on improved game theory

doi:10.1016/j.compeleceng.2020.106764

Computers & Electrical Engineering

Volume 87, October 2020, 106764

https://doi.org/10.1016/j.compeleceng.2020.106764 Get rights and content

Abstract

How to use the resources of the edge cloud more reasonably, reduce the energy consumption of machine equipment and ensure the shortest time for task completion are the challenges faced in cloud robot computation offloading research. In this paper, multiple heterogeneous cloud robot computing offloading problems are converted into game-type problems, and the computation-intensive tasks are divided to achieve partial offloading of tasks. An improved distributed game theory algorithm is designed to make each cloud robot's computation offloading strategy reaches the Nash equilibrium state, which maximizes the benefits of multiple participants, reduces the network load pressure of the central cloud, and reduces the transmission delay of computation offload. Simulation results show that the improved distributed game computation offload algorithm proposed enables cloud robots to reduce local computing energy consumption and shorten the average task completion time, greatly improving the edge cloud service quality.

Introduction

In recent years, the emergence of new applications such as face recognition, speech recognition, natural language processing, and virtual reality has caused the execution of related algorithms on embedded devices to consume a large amount of computing resources, while also meeting the user's requirements for low latency. Embedded devices are limited by their own computing, storage and other hardware resources when executing these new applications. Cloud robot [1] was proposed by Professor James Kuffner in 2010, which means that robots can offload computationally intensive tasks to the cloud, and use the rich computing resources of the central cloud to improve the quality of task completion. The traditional central cloud service model [2] has abundant computing resources, but it is easy to cause channel congestion and network delay. Edge computing [3] is an emerging computing model, which places computing, storage, bandwidth and other resources on the edge cloud near the device side to reduce transmission delay and bandwidth consumption. Cloud robots use computing offloading technology [4] to hand over some or all computing tasks to the cloud computing environment, which can help solve the problem of computing-intensive tasks for terminals with limited resources.

The cloud robot equipment divides computation-intensive tasks according to the offloading strategy, some tasks are left for local execution and some are uploaded to the edge cloud for cloud execution, as shown in Fig. 1. The edge cloud receives the request from the cloud robot, executes it, and returns the result to the cloud robot. This can reduce the cloud robot's energy consumption and improve task execution efficiency.

The rest of the paper is organized as follows: section 2 analyzes the research status of computing offloading, section 3 establishes a mathematical model for calculating offloading, section 4 designs improved game theory algorithms, section 5 provides simulation experiments and discussion of experimental results, section 6 is a summary of the paper.

Section snippets

Related work

In computation offloading research, based on partial task offloading, You et al. [5] studied a multi-user MEC system and devised an optimal strategy to control the amount of offloaded data and time/sub-channel allocation to minimize system energy consumption. Tao [6] proposed a solution to the problem of minimizing energy consumption of users, and obtained an optimal offloading strategy by using the Lagrange multiplier method. This strategy determined the optimal amount of offloaded data and

Computation offloading system model

Suppose there are N = {1, 2, 3, ......, n}cloud robot users, and each cloud robot runs a computationally intensive task. A task can be divided into M = {1, 2, 3, ......, m} subtasks, which can be partially unloaded. Taking a class of industrial robot [19] as an example, the task consists of five subtasks: sound preprocessing, feature extraction and classification, acoustic model loading, speech decoding and searching, and text output. Each node represents a subtask. Under the condition of order

The improved game model

Suppose N = {1, 2, 3, ......, n} is a collection of game participants,

$Q = [\begin{matrix} a_{11} & a_{12} & . . . & a_{1 n} \\ a_{21} & a_{22} & . . . & a_{2 n} \\ . . . & . . . & . . . & . . . \\ a_{m 1} & a_{m 2} & . . . & a_{m n} \end{matrix}]$ is the strategic space of the game, and I = {0, 1, 2, ......, I} is the game decision set.

The cost function is: $C_{n} = λ^{e} \sum_{m = 1}^{m} E_{m, n} + λ^{t} \sum_{m = 1}^{m} T_{m, n},$ where 0 ≤ λ^e ≤ 1,0 ≤ λ^t ≤ 1,λ^e + λ^t = 1.

To meet the needs of the cloud robot user, cloud robot users are allowed to select different weighting factors. For example, when the battery is low, a higher value of λ^e is assigned to save cloud

Simulation experiment

We used OpenStack, an open source cloud computing management platform, to build the edge cloud, and select the Medium virtual machine parameters in OpenStack to configure the parameters of the edge cloud. We set the kernel of the virtual machine to 2, with 12 GB of memory, a 30-GB hard drive, and a 5-GHz processor. Arduino is an open source prototyping platform that includes various versions of development boards and IDE kits. We used Arduino to create a cloud robot computing offload scenario.

Conclusion

In this paper, an improved game theory algorithm model is designed to assign different weight factors to cloud robots with different computing powers, so as to infer the best offload strategy and minimize the cost of computing offload.Simulation experiments analyze the cost of cloud robot systems with different calculation offload algorithms.The game theory algorithm proposed in this paper can reduce the system cost by 3.8% compared with the GA4CCO algorithm, and as the number of cloud robots

CRediT authorship contribution statement

Fei Xu: Conceptualization, Methodology, Validation. Weixia Yang: Formal analysis, Data curation. He Li: Software, Visualization.

Declaration of Competing Interest

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.

Acknowledgement

This paper is funded by the following funds: Natural Science Project of Shaanxi Education Department (18JK0399) ; fund of the State and Provincial Joint Engineering Lab. of Advanced Network, Monitoring and Control, China (GSYSJ2018006).

Fei Xu, is an associate professor at Xi'an Technological University, a master's tutor, and the director of the research and development center of the college. His research interests are edge computing, distributed systems, big data and adaptive middleware.

References (24)

Jinlin Guo et al.
Long collaborative computing offload algorithm based on optical fiber wireless network
Comput Eng Sci
(2019)
H Lu et al.
Low illumination underwater light field images reconstruction using deep convolutional neural networks
Future Gen Comput Syst
(2018)
J Kuffner J et al.
Space-filling trees: A new perspective on incremental search for motion planning
Yuezhi Zhou et al.
Near-end cloud computing: opportunities and challenges in the post-cloud computing era
Chin J Comput
(2019)
Weisong Shi et al.
Edge computing: current situation and prospect
Comput Res Dev
(2019)
Changsheng You et al.
Resource management for asynchronous mobile-edge computation offloading
C YOU et al.
Energy-efficient resource allocation for mobile-edge computation offloading
IEEE Trans Wirel Commun
(2017)
X TAO et al.
Performance guaranteed computation offloading for mobile-edge cloud computing
IEEE Wirel Commun Lett
(2017)
Y DAI et al.
Joint computation offloading and user association in multi-task mobile edge compu-ting
IEEE Trans Veh Technol
(2018)
Hongzhi Guo et al.
Collaborative computation offloading for multi-access edge computing over fiber-wireless network
IEEE Trans Veh Technol
(2018)

XIN J, KWONG K Y, YONG Y.Competitive cloud resource procurements via cloud brokerage Proceeding of the 5th IEEE...

X CHEN et al.

Efficient multi-user computation offloading for mobile-edge cloud computing

IEEE/ACM Trans Network

(2016)

Cited by (11)

CollOR: Distributed collaborative offloading and routing for tasks with QoS demands in multi-robot system
2024, Ad Hoc Networks
Multi-access edge computing (MEC) offers prospective opportunities for robots that have various computational tasks to execute in Industry 4.0, smart cities and many other fields. Computational offloading, as the key of this emerging paradigm, can leverage resourceful edge infrastructure nearby to provide computational support for these resource-limited robots. However, existing computational offloading schemes not only rarely consider that robots are geo-distributed, but also ignore the diversified quality of service (QoS) of robotic applications. Thus in this paper, the collaborative computational offloading problem for multi-robot system in MEC is studied, and a distributed scheme is proposed to jointly optimize offloading and routing for computational tasks of robots. First, we introduce the collaborative communication model of robots, construct the multi-criteria utility function related to both the QoS attributes of applications and the energy efficiency of robot, and formulate the joint offloading and routing problem. Afterwards, the QoS preference analysis mechanism is established to characterize the heterogeneous requirements of robotic applications. Next, by designing a novel collaborative robotics framework, we realize the distributed processing and adaptive strategy-making of multi-robots, and integrate the game-theoretic approach to propose a utility-aware collaborative computational offloading algorithm. Finally, simulation results validate that the proposed algorithm has significant performance improvement over the representative algorithms in different evaluation metrics, and can achieve stable and scalable performance in the scenario of the increasing demand for QoS and the growing scale of robots.
Dynamic computation offloading for ground and flying robots: Taxonomy, state of art, and future directions
2022, Computer Science Review
Citation Excerpt :
Compared to other predefined algorithms, the proposed solution was more stable in reaching optimal energy, delay, and path. Xu et al. [126] proposed an improved game theory algorithm to optimize both energy consumption and completion time. The idea consists of assigning each robot weighting factors to the completion time and energy consumption.
The collaboration between ground and flying robots can take advantage of the capabilities of each system to enhance the performance of robotic applications, including military, healthcare, disaster management, etc. Unfortunately, this cooperation is restricted by the physical constraint that all computation should be performed on robots. The concept of computation offloading has great potential to improve the performance of both ground and flying robots. Unfortunately, external conditions like the network conditions, the robot’s mobility, and the availability of the processing resources may lead to new challenges. Accordingly, these challenges should be addressed from different perspectives, like security, network communication, response time, and energy consumption. Recently, most computation offloading solutions aim to optimize further the robot’s energy consumption. Several research works are designed 1.) to satisfy the real-time requirements of robotic applications and 2.) to solve the trade-off between the energy consumed by computation and communication. To better understand these concepts, we present a comprehensive overview of the computation offloading process for ground and flying robots. We also devised a taxonomy explaining the factors affecting the offloading decision in a robotic scenario. The taxonomy presents guidelines to recognize the scope of research in offloading decisions that were designed for robots. Then, we discuss the state-of-the-art techniques of computation offloading from an architectural point of view, and we survey works related to offloading decisions for robots.
Binary task offloading strategy for cloud robots using improved game theory in cloud-edge collaboration
2024, Journal of Supercomputing
Real-Time Object Detection as a Service for UGVs Using Edge Cloud
2024, 2024 16th International Conference on COMmunication Systems and NETworkS, COMSNETS 2024
LATENCY AWARE INTELLIGENT TASK OFFLOADING SCHEME FOR EDGE-FOG-CLOUD COMPUTING – A REVIEW
2024, Informatics and Automation
DELOFF: Decentralized Learning-Based Task Offloading for Multi-UAVs in U2X-Assisted Heterogeneous Networks
2023, Drones

View all citing articles on Scopus

Weixia Yang, a master's student at Xi'an Technological University, mainly studies cloud computing and edge computing.

He Li, a master's student at Xi'an Technological University, mainly studies edge intelligence and deep learning.

View full text

Computation offloading algorithm for cloud robot based on improved game theory

Abstract

Introduction

Section snippets

Related work

Computation offloading system model

The improved game model

Simulation experiment

Conclusion

CRediT authorship contribution statement

Declaration of Competing Interest

Acknowledgement

Comput Eng Sci

Future Gen Comput Syst

Space-filling trees: A new perspective on incremental search for motion planning

Near-end cloud computing: opportunities and challenges in the post-cloud computing era

Chin J Comput

Edge computing: current situation and prospect

Comput Res Dev

Resource management for asynchronous mobile-edge computation offloading

Energy-efficient resource allocation for mobile-edge computation offloading

IEEE Trans Wirel Commun

Performance guaranteed computation offloading for mobile-edge cloud computing

IEEE Wirel Commun Lett

Joint computation offloading and user association in multi-task mobile edge compu-ting

IEEE Trans Veh Technol

Collaborative computation offloading for multi-access edge computing over fiber-wireless network

IEEE Trans Veh Technol

Efficient multi-user computation offloading for mobile-edge cloud computing

IEEE/ACM Trans Network