The motivation for this work lies in improving the performance of wheelchair guidance, even in the presence of noisy or ambiguous inputs from the user. The purpose of this paper is to implement the algorithmic procedures presented in two previous works that assess probabilistic decision-making in a real-time testing scenario by fusing spatial databases with temporal control inputs. Implementing this technique in addition to a cumulative weighted moving average allows for only minimal inputs to be provided within pre-defined decision-making regions. Consequently, a range of robotic applications can transition from fine to coarse control inputs, significantly reducing user/operator inputs while following their averaged path histories. A real-time demonstration of this technique is used in conjunction with a preview-based motion controller for path lock of a virtual wheelchair using actual joystick inputs.