| CPC G09B 19/003 (2013.01) [A63B 71/0622 (2013.01); A63B 2071/0641 (2013.01); A63B 2214/00 (2020.08); A63B 2230/625 (2013.01)] | 15 Claims |
|
1. A virtual reality-based task-oriented gait training system,
comprising:
a gait detection unit that detects a walking movement of a trainee;
a projector unit that projects virtual walking environment information for inducing a gait training of the trainee onto a projection surface corresponding to a ground; and
a control unit including a processor configured to execute instructions stored in a memory, the processor being operable to execute a gait training program to change the virtual walking environment information according to the walking movement of the trainee detected by the gait detection unit,
wherein
the virtual walking environment information includes a background constructing a virtual space, and a gait training object appearing in the virtual space, and
the gait training object includes at least one of a virtual obstacle for disturbing the walking movement of the trainee according to a preset level of difficulty and a gait inducing object for inducing a gait of the trainee with a predetermined stride length,
wherein the gait detection unit estimates a footprint pattern by recognizing central axes of both ankles of the trainee and both feet parts through analysis of data detected by a Lidar sensor, an RGB camera, or an IR depth sensor to classify and detect left and right feet, and detecting bending shapes of left and right ankles, and
the control unit compares the estimated footprint pattern with a normal gait pattern to discriminate an abnormal gait pattern,
wherein the estimating the footprint pattern includes using a footprint pattern estimation model machine-learned with training data constructed by databasing gait pattern images of a plurality of trainees, and
wherein the footprint pattern estimation model is trained by supervised learning using training data obtained by extracting and databasing foot pattern features detected from the gait pattern images of the plurality of trainees, so as to output the trainee's gait pattern as one of gait patterns including equinus gait, crouch gait, flatfoot gait, and bell gait.
|