Controlo Inteligente de um Exoesqueleto Flexível com Atuação Baseada em Têxteis

SmartExo aims to develop a prototype (TRL3) of a novel soft exoskeleton featuring a garment with a textile-based actuation (an exosuit) to actively assist the lower back and hip-knee joints, while ensuring maximal usability. The exosuit targets workers’ well-being and high productivity through Human- and task-sensitive control for augmenting workers’ force in physically demanding tasks, while preserving their natural dexterity to meet task requirements. This applied research is driven by the following intertwined objectives (OB), measurable through key performance indicators (KPIs) and milestones (M). OB1: To identify SmartExo requirements and the use case through a user-centred approach by creating an End-users and Stakeholders Panel(ESP). KPIs: number of members of ESP (1 industry, 2 workers, 1 ergonomist, 2 managers); documents stating requirements and the use case. They will be measured by M1 on 6th month. OB2: To design the exosuit’s fabric architecture and modelling the SMA actuator, resourcing Human MSK model, computational and CAD tools. KPIs: a more comprehensive constitutive model of SMA actuator than [5]; exosuit’s CAD design with approved requirements; document of fabrication guidelines (detailing SMA properties and its integration into textile); exosuit in TRL2. They will be measured by M2 on 17th month. OB3: To develop a novel active exosuit’s prototype. It aims at fabricate the garment with the SMA actuator from the guidelines, and integrate into exosuit the active heating and cooling methods, a safety layer, team-owned wearable sensor systems (WSS)[6,7], and power sourcing. KPIs: SMA-actuated exosuit with bandwidth>1Hz, power density>50W/kg, and force>50N when against a 5kg load, all beyond [4,8]; lightweight exosuit(<500g, stiffness/weight ratio>2 times[9]); efficient power consumption (< 70 Wh). KPIs are measurable by M4 on 26th month. OB4: To detect motor intention and physical effort (fatigue, metabolic cost, and ergonomic posture) during industrial tasks through AI tools fed by sEMG and kinematic data measured by WSS. KPIs: <5 sensors necessary for detection; more accurate tool than[2,10] for intention detection (>95%, delay<0.2s) of up to four tasks (walking, standing, squatting, and stooping); physical effort estimated with an RMSE<20%. These KPIs are measurable by M3 on 17th month. OB5: To develop Human- and task-sensitive control to personalize assistance beyond[2,10,11]. Novelty lies on developing an assist-as-needed control for adapting the assistance according to individual physical effort (human-in-the-loop) and intention; and, a multi-joint low-level controller, combining a SMA feedforward and a motion feedback loop to control heating and cooling temperatures. KPIs: controllers’ tunning <15min; controllers’ response compliant to worker’s needs and task’s restrictions (motion error<15o, heating and cooling rise time <1s [12,13]), all measurable by M5 on 32nd month. OB6: To conduct the proof-of-concept in the laboratory replicating a real use case of logistics. KPIs: new human motion datasets, exosuit rated with high usability (>4 in 5-Linkert scale), proper assistance (+30%joint torque, -20% physical effort). Aims TRL3 measurable by M6 on 36th month. OB7: To disseminate and transfer SmartExo’s results. KPIs: 2 conferences and 5 journal publications; 1 F.A.I.R. dataset; 7 SW modules, 1 prototype and its patent, advanced training (3MSc, 3 years of 2PhDs), and several outreach activities.