Power, sense, recognise
An implanted organ has to run without wires, read the body around it and decide when to act. Our three platform technologies address each of these, and together form a closed control loop.
Two-stage magnetic resonance wireless power
Deep implants are hard to power through the skin with a single pair of coils. We split the link into two resonant stages, with a relay under the skin.
- aExternal transmitter → subcutaneous relay → implanted receiver
- bDesigned for thicker tissue and changes in body position
- cPulse-switched valve held open by a permanent magnet, without continuous current
Most wireless medical power links couple one external coil directly to one implanted coil. Efficiency and stability fall quickly as tissue gets thicker or the patient changes position. Our architecture runs in series: an external transmitter powers a receiving coil fixed under the skin, and that coil acts as a relay, forwarding energy to the miniature coil of the device implanted at the bladder.
The two stages are designed and tuned as one matched system rather than two single-stage links added together. For the artificial-bladder valve, an electromagnet–permanent-magnet design delivers high torque at low power: a brief pulse switches the valve, and the permanent magnet holds the voiding position without continuous current. In company bench testing, this kept valve heating within 2 °C.
Flexible electronics for multimodal sensing
An artificial organ has to understand the body before it can act. Our sensing front end is a set of ultra-thin, flexible electrodes.
- aThin, stretchable electrodes that conform to soft tissue
- bBladder volume, pressure and temperature from one multi-source sensor
- cShared sensing base for the artificial-bladder, sphincter and kidney research lines
Unlike rigid metal probes, flexible electronics are thin films that bend and stretch. They conform to nerve, muscle or organ surfaces like a second skin and move with every pulse and deformation of the tissue, without opposing mechanical stress.
The same flexible front end receives several types of signal from different layers of the body at once. For the artificial bladder, a multi-source sensor is designed to monitor bladder volume, pressure and temperature together. These signals enter the recognition and control chain in parallel, so the device can both follow commands and track its own state — the precondition for closed-loop control. The front end is shared across our artificial-organ programmes rather than built for one device.
Real-time AI recognition
An end-to-end decoding pipeline turns ambiguous neural activity into a clear action the device can safely execute.
- aPreprocessing → feature extraction → pattern recognition
- bRuns on the device; no cloud round trip
- cCalibrated per patient and adapted as signals drift
Signals pass through preprocessing, feature extraction and pattern recognition. The pipeline is designed for low latency and clear discrimination, and it runs on the device rather than sending data to the cloud and waiting for a result.
In the artificial-bladder programme the trigger is the brain’s own voiding signal, read through a brain–computer interface (EEG). Neural signals are highly individual: two people’s intention to void looks different. The model is therefore calibrated to each patient at implantation or fitting, and continues to adapt as the tissue interface and signal quality drift over time. Once the intention to void is decoded, the system drives the implanted electromagnetic pump, so sensing, recognition and action happen in one loop inside the body.
- CNWireless electromagnetically controlled urination systemfiled
- CNUrination device and system with electromagnetic valvefiled
- CNSoftware for improving wireless energy-transfer efficiencysoftware copyright, registered
- CNSoftware for vital-sign-based monitoring and adjustmentsoftware copyright, registered