In minimally invasive surgery, surgeons often have to guide an endoscope at the same time as the procedure. This double burden can impair the stability of image guidance and increase physical strain, particularly during longer operations. A medical technology manufacturer has therefore developed a robotic assistance system that takes over camera guidance and can be controlled by voice command.

Voice-controlled robotics in the OR relies on a stable embedded platform

The aim of the system was to relieve surgeons and at the same time enable constant and reproducible image guidance. However, the technical implementation placed high demands on the underlying platform. In addition to low latency times for precise reactions, system stability in continuous operation, medical safety requirements, long-term hardware availability and future expandability also had to be taken into account.

An industrial embedded platform based on a 'Mini-ITX' motherboard with Intel processors, SSD memory and Linux system was used for the implementation. According to the company, system integrator Aaronn provided support in the selection, qualification and integration of the hardware into the existing robotics architecture. In particular, the focus was on the ratio of computing power, energy consumption and thermal stability.

The implementation was initially carried out using a proof of concept to test speech processing and system response. The platform was then gradually integrated into the robot system, interfaces were adapted and tests were carried out under realistic conditions. A particular focus was placed on reproducible system states in order to facilitate subsequent validation and certification processes.

Latencies under 100 ms

According to the company, the system processes voice commands with latencies of less than 100 ms. Longer test phases also showed stable continuous operation without any relevant failures. The integration was therefore possible without any fundamental new hardware developments, while the platform was also designed for long-term medical application scenarios.

The project makes it clear that computing power is not the only decisive factor in medical robotics systems. Integration capability, validation, reproducible processes and the long-term stability of the platform used are often more important. Particularly in sensitive areas such as endoscopy, the system architecture is therefore crucial to its suitability for practical use.

With a view to future developments, the selected platform should also enable adjustments to software, sensor technology or computing power without the need for fundamental changes to the system architecture.