 Hi, I am Ours Kaihan. Today I will present a variable haptic guidance system based on skin stretch around the waist for visually impaired runners. We have developed a guidance system as a belt, and this system can be divided into two parts, which are environmental sensing and haptic feedback. Environmental sensing part consists of a microcontroller and RGBD camera. For haptic feedback part includes two servo motors and polysystem. In order to grip the skin, the latex straps are used. All these pieces are attached to the fabric belt. There are three types of feedback are given by the prototype. If the running lane curves right, both servo rotate counterclockwise. The left side of the waist is stretched forward, and the right side of the waist is stretched backward. If the users should turn left, both servo rotate clockwise. The left side is pulled backward, while the right side of the waist is pulled forward. To give the slowdown and stop feedback, the left servo rotate clockwise, and the right servo rotate counterclockwise. The user is expected to feel that both sides of the waist are stretched backwards. In order to evaluate usability and overall effectiveness of the prototype, two experiments were conducted. These experiments were completed with four participants. Three of them didn't have any disabilities, and one participant was visually impaired. In the static test, the participants were asked to identify the given haptic feedback on their waist while standing still. Right after receiving a random feedback signal, the participants indicated the direction and the level of the feedback on an estimation chart shown in the figure. A training session was performed before the experimental trail. In the dynamic test, the system was evaluated by asking the participants to run 400 meters with the help of the prototype. Able-bodied participants were blindfolded. Number of lane violations, number of verbal warnings, and the total time of the run were noted by the experiment coordinator. In order to compare the results, reference run without haptic guidance were also performed. They were navigated with only the verbal warnings. The static test was performed twice by each participant. The accuracy in the second trial is higher than the accuracy in the first trial. Higher accuracy in the second trial indicate that the device could be learned quickly. A confusion matrix is expected to be diagonal for an optimum feedback system. Despite the second test result clearly better than the first one, over the 70% of the error in the second test are related to confusion in the direction of the feedback. However, as the training decreased the total number of errors considerably, further usage of the device might lower the directional confusion as well. In the dynamic test, the visually impaired participant and the able-bodied participants should be evaluated separately. Since the able-bodied participants were in an unfamiliar situation, the feeling of the fear and insecurity directly affected their performance. The number of lane violations and the verbal directional warnings decreased for the visually impaired participant when the device was turned on. On the contrary, an increase in the total time of the run is observed. The decrease in the running speed may be due to the trying to understand the feedback of the device. Regarding the able-bodied subjects, the run performance with the proposed device had less lane violations and directional warnings than the reference run. To conclude, an all-variable assistive device for visually impaired runners has been designed, developed and tested. There are many studies and devices in the literature and in the market for the guidance of the visually impaired persons. However, there are only few studies targeting the visually impaired persons who do supports and the feedback system of these studies are solely on vibration and sound. In this study, we propose a device that uses second-stage feedback and designed as a belt for visually impaired runners. Thank you for your attention.