Abstract—More than 62,000 Australians were reported suffered from Traumatic Brain Injury (TBI), Spinal Cord Injury (SCI) and Cerebrovascular Accident (CVA) or stroke in 2011. These injuries and accidents lead to physical disability that yields in limitation for performing a person’s daily life activities. To overcome such limitations, physical rehabilitation is conducted which requires one-to-one attention that creates a shortage in therapists and lead to high cost. In this paper, a development of an effective augmented reality (AR) based upper limb rehabilitation system with low cost is presented. Our development aims to close the gap in shortage of therapists, high health care cost of TBI, SCI and stroke. The proposed system can be used at home as well as in rehabilitation centers, units, and hospitals with minimum therapist supervision. It consists of two modules: AR based rehabilitation exercises module and real-time active muscle module. The first module aims to increase the upper limb range of motion via reaching exercises, and strengthen the associated muscles. In second module, the patient’s EMG signals were used as an input to monitor the muscle performance in real time during training. Our development had tested with 10 healthy subjects and had demonstrated in Port Kembla Rehabilitation Hospital. I. INTRODUCTION EUROTRAUMA such as Traumatic Brain Injury (TBI), Spinal Cord Injury (SCI) or Cerebrovascular Accident (CVA) or stroke is the main reason of physical disability. It reported that 2,493 cases of TBI, 362 cases of SCI and 60,000 stroke cases were occurred in 2011[1]. The lifetime costs of TBI and SCI were estimated to be $10.5 billion and the cost of stroke has reported around 2.14 billion per year in Australia according to WAIMR[1]. People who suffer from TBI, SCI or stroke are faced with loss of control over one side of the body generally. Thus, patients cannot perform the daily live activities by themselves and this impact them and their families’ quality of life deeply. However, the studies had proven that performing of repetitive tasks and task-orientated activities can improve this type of motor impairment[2]. Hence, a lot of upper limb rehabilitation systems had researched and developed for restoration of lost functions. Such developments include robotic approach (end-effector based [3, 4] and exoskeleton based [5-7]), virtual reality (VR) based approach and augmented reality (AR) based approach. Generally, robotic systems aim to train for severe impairments and classified as expensive assistive device while VR and AR based systems aim for minor impairment or later stage of rehabilitation training at low cost. The latter approaches provide with Yee Mon Aung and Adel Al-Jumaily are with School of Electrical, Mechanical and Mechatronic Systems, University of Technology Sydney, Australia, 15 Broadway, Ultimo, NSW 2007 (yee.m.aung@student.uts.edu.au; adel.al.jumaily@uts.edu.au)). better encouragement and motivation as these systems employ game based exercise as a training platform. Some researchers developed the combination of robotic and VR based approach as in [8]. Research studies had confirmed that the embedded of VR in rehabilitation system provides positive results [9, 10]which reflect on several developments [11, 12]. Stand-alone VR based rehabilitation system integrated with biofeedback system also can be found in [13]. Although VR based developments have proven with positive results, additional attachment of tracking device to the patient, their bulkiness and total immersive in virtual world are inconvenient and dangerous for patients especially if the patient is a child. Therefore, Augmented Reality (AR) based rehabilitation exercises have been developed for better and safer interactive environment. Augmented reality is the combination of real world and virtual world that enhance the user perception of reality. The user can view the computer generated virtual environment that is overlaid on top of real environment. As far as AR based rehabilitation system is concerned, J. W. Burke et al. [14, 15] have developed several exercises for upper-limb stroke patients. His development tracks the marker that is a defined color object to interact with virtual display on the display screen. His development aimed to obtain back the patient’s motor functions such as grasping, reaching, lifting, releasing and cognitive skills. Another AR based upper limb rehabilitation exercise is AR-REHAB. It was developed by Atif Alamri et al. [16] for post stork patient rehabilitation. The developed system was aimed to improve the patient’s arm reaching and hand grasping ability.AR drink, AR dance and AR fold were developed to improve the coordination of stroke patients in [17]. The system was developed to train the patients’ upper limb for daily life activities such as drinking, dancing and folding via virtual objects. Dunne et al. [18]invented the rehabilitation system with multi-touch display for the children with Cerebral Palsy (CP). One of the main features of this system was tracking the trunk position of patient and prevent from compensatory movement. The researchers in [19] developed two augmented environments (AE) for paediatric rehabilitation to improve the motor control via music playing AE and topographical orientation training to relearn the community mobility skills through decision making in AE. Another AR based rehabilitation musical system was developed in[20]. It was developed for children with CP to rehabilitate the arm movement via computer assisted music therapy. The rehabilitation purpose of this intervention is relearning cognitive, motor, psychological, social activities skill. AR based hand rehabilitation system was invented and reported in [21]. In this development, virtual piano as a hand rehabilitation therapy with self- designed data glove that detect the flexing movements of AR based Upper Limb Rehabilitation System Yee Mon Aung, Student Member, IEEE and Adel Al-Jumaily, Member, IEEE N The Fourth IEEE RAS/EMBS International Conference on Biomedical Robotics and Biomechatronics Roma, Italy. June 24-27, 2012 978-1-4577-1198-5/12/$26.00 ©2012 IEEE 213