Many years ago, Tianjin People's Hospital welcomed the first patient of the stroke neural rehabilitation robot project - Aunt Shi, a 64 year old retired teacher. At that time, it was difficult for Aunt Shi, who was deeply troubled by stroke, to hold a pen. "My mother has been teaching and nurturing on the podium all her life, but now that she's retired, she can't even hold a pen. She can't accept it, and her mood is very low every day." Zhang Hao, an assistant researcher and postdoctoral fellow at Tianjin University, still remembers these words from her aunt's children. At the second Postdoctoral Innovation and Entrepreneurship Competition held in Yantai, Shandong not long ago, Zhang Hao participated in the leading track with the project "Neurostroke Rehabilitation Robot Based on EEG Collection, Processing and Analysis". Zhang Hao is the project leader, and his team is guided by clinical stroke rehabilitation needs to design and develop an artificial neural rehabilitation robot system. At present, this achievement has been clinically studied in multiple tertiary hospitals. "Raise your right hand", "move your fingers", "take a small step forward"... At the Rehabilitation Department of Tianjin People's Hospital, 63 year old stroke patient Wang Baozhan (pseudonym) wore a EEG hat and attempted to use his mind to "control" his limbs, showing slight dullness. Zhang Hao told China Youth Daily and China Youth Network reporters, "The limb dysfunction caused by stroke seriously affects the quality of life of patients, causing a huge burden on individuals, families, and society." In his view, in recent years, the neural rehabilitation robot system developed based on motion imagination brain computer interface technology can achieve direct interaction between the brain and muscle bone control devices, induce the recombination, repair, and enhancement of neural network connections, The reconstruction and rehabilitation of motor function have positive significance. However, the complex interaction information at the neural system level faced by high-performance human-computer interaction makes it difficult for existing neural rehabilitation robot systems to efficiently decode neural information and accurately evaluate the rehabilitation process. The stroke neural rehabilitation robot developed by Zhang Hao's team uses brain computer interface technology to detect real-time signals emitted by the brain through a EEG cap placed on the scalp, achieving brain machine muscle collaborative interaction. It repeatedly trains and strengthens the excitation transmission pathway from the brain to muscle groups, effectively promoting brain nerve function remodeling and helping patients recover their motor function. Zhang Lei, a core member of the stroke neuro rehabilitation robot project and a postdoctoral fellow of Tianjin University, said that compared with the traditional treatment methods focusing on limb training such as acupuncture and moxibustion, massage and physiotherapy, the stroke neuro rehabilitation robot can train the brain and limbs at the same time, help the rehabilitation of neural plasticity, and greatly reduce the treatment time. Based on this training mode, Aunt Shi can write her name neatly after 5 months of training. Zhang Hao introduced that there are still urgent problems that need to be solved for stroke neural rehabilitation robots. For the treatment of stroke patients, invasive methods are mainly used in foreign countries, while non-invasive methods are mainly used in China. "No craniotomy or invasive trauma surgery is required, just wear a EEG cap.". In Zhang Hao's view, this also brings new problems: non-invasive means require neural signal transduction through structures such as brain tissue, cerebrospinal fluid, skull, and scalp, resulting in poor quality. How to improve the accuracy of scalp EEG signals to an invasive level is still being explored by his team. He and his team are passionate about entrepreneurship