Multiple sensors group nearby people into zones so a motorized chair can match crowd movement and maintain safe spacing for collision avoidance.
Real-time feedback and selective CNS/PNS stimulation help restore smoother walking while reducing parasite movements after neurological injury.
Collision-based digital models verify dental restoration insertion paths accurately, avoiding costly and time-consuming physical prototypes.
A representative glucose curve fitted to a PI control model yields digital biomarkers that distinguish healthy, prediabetic, and diabetic glycemic control.
Segmented virtual boundaries and haptic tip forces help teleoperated repositionable arms avoid collisions earlier without overly restricting workspace.
Hybrid automata update patient predictive models from real-time inputs, improving insulin dosing accuracy and glucose safety under changing conditions.
Digitized spinal attachment points let a computer define rod bend location, angle, and rotation, reducing surgical time and rebending fatigue.
A tissue-mimicking 3D hydrogel phantom enables direct TTField shape and intensity measurement, improving tumor targeting beyond simulation alone.
Uses max-margin temporal transduction to detect health-state changes and change-points from time-series data with limited labels.
Robotic arm kinematics and setup data seed anatomy registration, aligning intra-operative and pre-op images without patient pads.
Real-time ML guidance identifies anatomy and procedural steps during minimally invasive surgery to reduce misidentification and surgical mistakes.
Chronology-constrained transductive SVMs use unlabeled time-series data to detect healthy-to-unhealthy transitions and their timing.
Mobile images and patient data generate parametric CAD for 3D-printed orthotics with accurate fit, variable density, and cushioned support.
Patient-specific bend parameters guide spinal rod angle, rotation, and position to cut surgical time and reduce metal fatigue.
A multi-model predictive controller switches fast and slow insulin kinetics with Kalman filtering to keep glucose levels stable.
Successive 3D-shaped nose instruments guide cartilage growth to achieve a target shape with more comfort and predictability than surgery or rigid splints.
Real-time EEG source imaging and continuous pursuit training improve noninvasive robotic arm control despite weak EEG signal quality.