Harnessing Industrial Internet of Things and Autonomous Robotics for Resilient Production Networks in Advanced Digital Factories

Manufacturing systems are increasingly expected to absorb disruption rather than merely avoid it, yet the two technological pillars most often invoked to deliver that capacity   the Industrial Internet of Things and autonomous robotics   have been developed and evaluated largely in isolation. This article addresses a specific empirical gap : the absence of an integrated account of how IIoT-mediated perception and autonomous robotic actuation jointly produce resilience at the level of the production network, supported by measurable evidence. Through a systematic review and framework synthesis of the relevant literature, the study develops a four-capability resilience architecture   sensing, deciding, acting and recovering   and formalizes it as a measurable index combining performance loss, settling time and capability weights. The analysis yields four findings. First, the three contributing literatures are structurally adjacent rather than connected : architectural work is platform-centric, robotics work is cell-centric, and resilience work is metric-centric, so none of the three alone explains network-level behaviour under disruption. Second, the resilience metrics available in the literature are configuration-dependent rather than absolute, since measured resilience depends on system configuration, levels of redundancy and flexibility, and buffer capacities. Third, the robotics literature identifies certification of auxiliary sensors, large safety zones and closed control architectures as concrete deployment bottlenecks, which converts a design problem into a compliance and integration problem. Fourth, redundancy and graceful degradation emerge as the most consistently evidenced mechanisms, whereas proactive readiness strategies remain the least developed phase of the resilience cycle. The article concludes that the empirical gap is closable through a common experimental protocol built on configuration-parameterized resilience metrics, and formulates five propositions to guide it.