- 01DVL taps movement data, Reflex Engine, RRAM for robotics.
- 02Target balance, coordination, recovery.
- 03Data library spans gait, posture; wearables/Video AI for tests.
dorsaVi (ASX: DVL) has begun engaging with robotics developers to explore how its established movement data library, bespoke capture capabilities, and computing technologies could be applied to humanoid robotics.
The initial focus is balance, coordination, and recovery—areas where dorsaVi believes real-world movement recordings can help define how machines detect instability and respond across multiple joints.
The company is combining more than a decade of movement data with its exclusively licensed Reflex Engine neuromorphic intellectual property and developing resistive random-access memory (RRAM) hardware as part of a broader physical AI strategy.
Discussions with robotics developers will seek to identify application requirements and potential technical evaluations before dorsaVi progresses towards any application-specific engineering or joint development.
Movement Data Meets Robotics
dorsaVi’s existing library spans clinical, workplace, and sporting environments and includes gait, changes of direction, postural control, recovery from instability, loading, and movement affected by injury or restricted motion.
The company can supplement that resource with bespoke recordings using wearable sensors and Video AI systems, allowing specific tasks, body segments, and operating conditions to be captured for potential partner requirements.
Its proposed robotics applications include analysing how coordinated movement changes through turns, posture shifts, uneven loading, and slips, with selected recordings requiring preparation and adaptation to each robot’s physical structure, actuators, and control architecture.
The current internal review will assess the coverage, fidelity, sampling characteristics, and consistency of the data, as well as the preparation, structuring, and labelling needed before selected datasets could support movement modelling or hardware evaluation.
Local Sensor Responses
The Reflex Engine is designed to support rapid local responses to sensor inputs, giving dorsaVi a potential processing layer for balance and coordination tasks without relying solely on instructions from a central processor.
dorsaVi’s RRAM development program is meanwhile progressing with two international research institutes, with 180 nanometres (nm) validation intended to inform development towards a future 22nm platform.
The company plans to investigate whether the analogue characteristics of RRAM materials can support selected compute-in-memory functions, including neuromorphic processing and physical artificial neural networks.
“For more than a decade, dorsaVi has collected detailed movement data in clinics, workplaces, and sporting environments that already supports occupational health and safety studies, and we are now reaching out to robotics companies to explore its potential in a new market,” group chief executive officer Mathew Regan said.
“We see a clear opportunity to extend the value of our existing movement data and capture capabilities while informing the development of our neuromorphic and RRAM technologies.”
Candidate Workload Definition
dorsaVi intends to use feedback from robotics developers and its movement data review to define candidate workloads covering how fast, how granularly, and how locally a balance correction must occur.
Those workload definitions could then become reference requirements for assessing RRAM materials and evaluating whether the Reflex Engine can address selected balance and coordination tasks.
Whether suitable RRAM materials can ultimately be identified and developed into devices remains undetermined, while integration into robotic systems would require application-specific engineering and validation.
Subject to the review findings, the next steps are expected to include preparing selected datasets, defining a Reflex Engine workload for a balance application and engaging robotics developers on potential joint evaluation.
Get the wire before the market opens.
The ASX small-cap stories that matter, filed before 9am AEST. Curated by the Small Caps desk.