Sistem za posnemanje spotikanja in pomoč pri zamahu noge pri hoji po tekočem traku

TOMC, MATEJ (avtor), Mihelj, Matjaž (mentor), Matjačič, Zlatko (komentor)

Padci predstavljajo visoko tveganje za poškodbo, zlasti pri starejših. Pacienti, ki okrevajo po možganski kapi, imajo slabše dinamično ravnotežje, zato so za padce dodatno dovzetni. Poleg tega imajo zaradi oslabelosti mišic in spastičnosti v svojem vzorcu hoje številne nepravilnosti. Vzrok velikega deleža padcev so spotiki. V procesu rehabilitacije je pomembno, da iz pacientovega cikla hoje eliminiramo deviacije od zdravega cikla hoje, katerih prisotnost povečuje tveganje za spotik in padec. Ker se spotikanju v realnem okolju ne moremo izogniti, je pomembno, da je v rehabilitacijski proces vključen tudi trening povrnitve dinamičnega ravnotežja. Skozi trening se pacient nauči strategij, ki mu omogočajo, da se kljub spotiku izogne padcu. V modernih programih rehabilitacije si veliko pomagamo z roboti. Naš cilj je razviti napravo za rehabilitacijo spodnjih ekstremitet pri hoji po tekočem traku, ki deluje po zgledu tetivnih robotov. Naprava bo modularna, cenovno dostopna in varna, saj razen tekočega traku ne bo vsebovala aktivnih elementov. V tem delu predstavljamo razvoj dveh podsistemov naprave. Prvi je sistem za posnemanje spotikanja, drugi pa za pomoč pri zamahu noge pri hoji po tekočem traku. Razvili smo napravo za posnemanje spotikanja in izdelali njen prototip. Naprava omogoča izvajanje perturbacij na nogo med hojo po tekočem traku. Perturbacijo predstavlja zaustavitev gibanja gležnja v anteriorni smeri za kratek časovni interval. Delovanje naprave smo evalvirali na enem zdravem prostovoljcu. Prostovoljec je hodil po instrumentacijskem tekočem traku, s katerim smo merili reakcijsko silo podlage v vseh smereh ter mesto prijemališča reakcijske sile podlage na tekočem traku. Na izbranih anatomskih mestih telesa prostovoljca so bili nameščeni odsevni markerji, s pomočjo katerih smo s sistemom kamer za zajem gibanja posneli kinematiko segmentov telesa. Prostovoljec je hodil pri treh počasnih hitrostih hoje. Rezultati kažejo, da algoritem za proženje perturbacij razpoznava cikle hoje ter robustno in ponovljivo proži perturbacijo v izbranem trenutku znotraj cikla hoje. Meritve sile podlage in prijemališča sile podlage kažejo, da je ravnotežni odziv po perturbaciji podoben ravnotežnim odzivom po spotiku, ki jih najdemo v literaturi. Posnetek kinematike segmentov telesa nakazuje na to, da lahko z ustrezno nastavitvijo dolžine perturbacije in hitrosti tekočega traku ponovljivo izzovemo strategije za povrnitev ravnotežja, ki jih opisuje literatura. Sistem za pomoč pri zamahu noge pri hoji po tekočem traku smo razvili v simulacijskem okolju. Izdelali smo dinamični model noge, ki posnema delovanje trisegmentnega ravninskega serijskega robotskega mehanizma. Dinamiko noge v zamahu smo modelirali z Newton-Eulerjevo metodo. Oslabelost mišic smo modelirali z zmanjšanimi navori v sklepih, spastičnost pa kot povečani mišični tonus, ki ustvarja navore v sklepih, ki zavirajo gibanje. Zamah hemiparetične noge smo želeli popraviti z uporabo elastičnih vrvi. V elastičnih vrveh se v fazi opore shrani energija, ki jo v sistem vnaša tekoči trak, v fazi zamaha pa vrvi pomagajo, da hemiparetična noga izvede zamah, katerega kinematika je čim bolj podobna zamahu pri zdravi hoji. Za vsako elastično vrv je mogoče izbrati trenutek začetka napenjanja in trenutek prenehanja delovanja vrvi. Simulirali smo vodenje sistema z genetskim algoritmom ter z algoritmom vodenja z iterativnim učenjem (ILC). Oba algoritma omogočata zadovoljivo vodenje napenjanja in delovanja elastičnih vrvi v simulacijskem okolju. V obeh primerih lahko z uporabo elastičnih vrvi dosežemo, da postane zamah hemiparetične noge v ključnih parametrih bolj podoben zamahu zdrave noge. Zaradi lažje implementacije bomo v realnem sistemu za vodenje uporabili algoritem vodenja z iterativnim učenjem. S pomočjo simulacije smo določiti tudi ostale proste parametre pri načrtovanju sistema. Iz evalvacije prvega podsistema sklepamo, da je človekov odziv na perturbacijo, ki jo izvedemo z našo napravo za posnemanje spotikanja, dovolj podoben odzivu na spotik ob fizično oviro, da bi bilo napravo smiselno uporabiti v postopku rehabilitacije. Rezultati simulacijske študije sistema za pomoč pri zamahu noge pri hoji po tekočem traku kažejo, da je sistem pripravljen za implementacijo in primeren za uporabo v rehabilitaciji hoje pacientov, ki okrevajo po možganski kapi.

• Jezik: Slovenski jezik
• Ključne besede: padec, rehabilitacija, možganska kap, tetivni robot, spotikanje, cikel hoje, simulacija
• Vrsta gradiva: Magistrsko delo/naloga
• Organizacija: FE - Fakulteta za elektrotehniko
• Leto izida: 2020
• PID: 20.500.12556/RUL-118881
• Datum objave v RUL: 01.09.2020

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: System for tripping emulation and swing assistance during treadmill walking

Izvleček

Falls are a major health hazard. The risk of falls is higher for the elderly as well as stroke survivors, whose dynamic balancing capabilities are impaired. Stroke survivors also suffer from muscle weakness and spasticity, which result in a gait, that differs significantly from the gait of the able-bodied. A large part of falls is caused by trips. The goal of the rehabilitation process is to reduce gait deviations that increase the risk of tripping and falling. It is also important for the rehabilitation process to include dynamic balancing training. This training helps patients learn how to employ trip recovery strategies in the events when tripping unexpectedly occurs. Modern rehabilitation processes are often supported by rehabilitation robots. Our goal was to develop a lower limb rehabilitation device for treadmill walking, inspired by tendon-driven robots. The device should be modular, inexpensive and safe. The only active part of the system is going to be a treadmill. Development of two subsystems that will make up this device is presented. The first one is a system for tripping emulation and the second one is a system for swing assistance during treadmill walking. A device for tripping emulation during treadmill walking was developed and prototyped. The device can create gait perturbations by preventing the ankle from moving in the anterior direction for the duration of a short interval of time. To evaluate the device a tripping experiment was conducted on one healthy subject. The subject walked on an instrumented treadmill at three slow walking speeds. Measurement of ground reaction force and center of pressure were taken. Measurements of kinematics of body segments were made using a full-body set of reflective markers and a motion capture system. Results indicate that the algorithm that triggers perturbations is able to successfully segment gait cycles into basic gait phases. Perturbations are triggered in the correct timeframe within the swing phase in a repeatable and robust manner. Measurements of center of pressure and ground reaction force indicate that the perturbation induced balancing response is similar to balancing responses after tripping described in literature. The system for swing assistance during treadmill walking was developed using a simulation software. A human leg dynamic model was developed. The leg is modelled like a three-link planar serial robot using the Newton-Euler method. Muscle weakness was modelled with a decrease in joint torques. Spasticity was modelled like increased muscle tone, that creates torques which oppose joint movement. The goal of the system is to assist with the movement of the hemiparetic limb in the swing phase. This is done using elastic ropes. Elastic ropes can store energy by extending while the leg is moving backwards on the treadmill, and release this energy in the swing phase, causing the movement of the hemiparetic leg to better match the movement of an able-bodied person’s leg in swing phase. The time when an elastic rope starts extending and the time when the assistance of a rope is cut off can be controlled for each individual rope. Two control algorithms were tested: a genetic algorithm and an iterative learning control algorithm. Both control algorithms produce satisfactory results, improving the key parameters of the hemiparetic swing when elastic ropes are used for assistance. Due to easier future implementation in a real-time computing environment, the iterative learning control algorithm was chosen. Other design parameters of the system were also tested and chosen using this simulation. From the evaluation of the first subsystem we conclude that a person’s balancing response, induced by the device for tripping emulation, closely resembles the balancing response which occurs after tripping over a physical obstacle. Therefore, the device has potential for use in rehabilitation. Results of the simulation of the system for swing assistance during treadmill walking indicate that the system is ready for implementation and can be used in rehabilitation of stroke survivors.

• Ključne besede: fall, rehabilitation, stroke, tendon-driven robot, trip, gait cycle, simulation