Human-Agent-Robot Teamwork (HART) Over FiWi-Based Tactile Internet Infrastructures

Human-Agent-Robot Teamwork (HART) Over FiWi-Based Tactile Internet Infrastructures

Mahfuzulhoq Chowdhury (Institut National de la Recherche Scientifique (INRS), Canada & University of Quebec, Canada) and Martin Maier (Institut National de la RechercheScientifique (INRS), Canada)
Copyright: © 2021 |Pages: 15
DOI: 10.4018/978-1-7998-3479-3.ch003
OnDemand PDF Download:
No Current Special Offers


To facilitate making the human-machine co-activity-based human-agent-robot teamwork (HART) task execution process more efficient, this chapter first discusses related work and open challenges for latency-sensitive HART task. To speed up the HART task execution, this chapter next presents a latency sensitive HART task migration scheme for efficiently orchestrating tasks among human mobile users (MUs), central and decentralized computational agents (cloud/cloudlets), and robots across converged FiWi network infrastructures. Moreover, this chapter describes a bandwidth allocation scheme that allocates timeslots to MUs' broadband and task migration traffic at the same time. Furthermore, this chapter presents performance evaluation results of the proposed scheme. Importantly, this chapter compares the performance of the proposed task migration scheme with traditional schemes. This chapter is finally concluded by summarizing important findings and outlining open research issues for HART task coordination over FiWi-enhanced tactile internet infrastructures.
Chapter Preview


In the current digital Darwinism era, technologies are evolving faster than the business organization can naturally adapt (Daugherty, 2018). In the past, business executives focused on using machines to automate specific workflow processes. Traditionally, these processes are linear, stepwise, sequential, and repeatable. But performance gains from that traditional approach have recently been leveling off, as companies wring the last bits of efficiencies from mechanistic automation. The traditional approach (either human-only activity or machine-only activity) has been to view humans and machines as rivals, each side fighting for the other's job. The truth is that companies can achieve significant boosts in performance, when machines and humans work together as allies, not adversaries, to capitalize on each other's complementary strengths. For instance, performing routine tasks and detecting hidden patterns in an image can be easy for machines, whereas dealing with unsatisfied customers can be easy for humans. To achieve higher productivity gains than traditional systems, currently leading firms in many industries are re-imagining their business processes to be more flexible, faster, and adaptable to customer behavior, preferences, and needs of their workers in a given moment. This trend of exploiting the adaptive system for symbiotic collaboration between humans and machines is unlocking what is known as the third wave of business process transformation. This collaboration allows weak workers to transfer tedious tasks to powerful machines and by enabling them to perform their work faster and more efficiently through expert guidance, advice, and support from artificial intelligence (AI) based systems. Hence, to realize beneficial human-machine coactivity in different business processes, the fundamental question that naturally arises is how to enable real-time collaboration and communication between humans and machines. To answer this question, some of the key technological terms are described below in greater detail:

To unleash the full potential of many real-time cyber-physical systems (CPSs) that harness human-machine interaction, including virtual and augmented reality, an extremely low round-trip latency of 1-10 milliseconds is required to match human interaction with the environment. This vision of ultra-low latency communications gives rise to the so-called Tactile Internet, which has recently emerged to steer/control virtual and physical objects (e.g., remote-controlled robots) and to transmit touch and actuation in real-time (Maier, 2016). Key Tactile Internet applications include real-time gaming, wireless controlled exoskeletons, remote surgery, elderly/disabled people care, driver-assistance systems to support humans in arriving safely and comfortably at their destinations. By enabling ultra-low latency communications between humans and machines in conjunction with carrier-grade robustness and availability, the Tactile Internet represents a paradigm shift from traditional wired and mobile Internet based content-delivery to labor-delivery networks via tactile/haptic devices. At the nexus of computerization, automation, and robotization lies the emerging Tactile Internet, which will be centered around human-to-machine/robot (H2M/R) communications by providing the medium for transporting haptic (i.e., touch and actuation) and non-haptic (data, video, and audio) traffic in real-time (Maier, 2018). By offering highly reliable and responsive network connectivity for real-time human-machine interaction centric applications, the Tactile Internet holds great promise to have a profound socio-economic impact on our everyday life, ranging from healthcare (e.g., remote robotic surgery), transport (e.g., driver assistance system) to entertainment (e.g., online gaming) and manufacturing industry. The long-term goal of the Tactile Internet is to enable emerging products and services that complement humans rather than substitute for them. Recently, Stanford University launches its inaugural report on “Artificial Intelligence and Life in 2030,” in which they predicted that within next 10-15 years a vast amount of intelligent human-aware systems will be developed. Unlike the Internet of Things (IoT) without any human involvement in its underlying machine-to-machine (M2M) communications, the Tactile Internet involves the inherent human-in-the-loop (HITL) nature of haptic interaction and thus allows for a human-machine cooperative design approach towards creating and consuming novel immersive experiences via the Internet.

Key Terms in this Chapter

Human-Agent-Robot Teamwork (HART): Human-agent-robot teamwork (HART) denotes the task and resource coordination framework for executing a requested task in a collaborative setting, in which human mobile device transfers the task request and receives the task result from collaborative machines (robot/cloud agent) after processing. Whereas, the robot and cloud agent process the physical (movement towards task location and image capturing using camera) and cognitive sub-task (face detection from captured image), respectively. Note that, in this work the robot uploads the physical sub-task output data size (input of cognitive sub-task) to cloud agent for processing. Hence, the cloud server/agent transfers the processed cognitive sub-task results to human MU’s device after processing.

Task Migration: Contemporary mobile devices offload large amounts of computationally intensive tasks to resource-rich cloud servers for processing. Task migration broadens the scope of conventional computation task offloading by not only transferring the task from an MU to cloud servers/surrogates but also from one cloud server/surrogate to another one for execution.

Mobile-Edge Computing: Mobile-edge computing (MEC) technology offers decentralized cloud computing services (e.g., computing storages) to mobile users’ devices for executing their computation-intensive tasks/applications by setting up local cloud servers/cloudlet/fog servers at the network edge (WiFi/cellular base station) in close proximity to mobile users through a process named as computation offloading.

Predicted Task Execution Time: Predicted task execution time denotes the time interval between task request assignment by task coordinator to suitable actors (robot/cloud server) and task result reception by human MU’s device.

Tactile Internet: The tactile internet is defined as a very low-latency communication system that ensures very low round-trip delay along with high availability, reliability, and security for real-time human-machine interaction-centric applications execution.

Fiber-Wireless Enhanced Networks: Traditional FiWi access networks are based on integration of both optical fiber (Ethernet passive optical network or EPON) technologies (at the wired backhaul) and wireless (wireless local area network or WLAN) Ethernet technologies (at the wireless frond-end), which are then integrated with their cellular counterparts, namely, 4G Long Term Evolution Advanced (LTE-A), to give rise to FiWi enhanced LTE-A heterogeneous networks (HetNets).

Complete Chapter List

Search this Book: