Quill builds small inspection platforms with compliant arms and a vision stack that does not hallucinate cracks. Aerospace and pharma are the current manners.
They commission work packages for perception, certification evidence, and the odd custom end-effector.
Their design philosophy is that an inspection robot's most important behaviour is what it does when it is uncertain. A platform that reports 'I cannot see this properly' is useful; one that returns a confident wrong answer destroys the inspection regime it was meant to support. This has made their perception work more conservative and their customers considerably calmer.
The commercial obstacle is evidence rather than capability. Regulated industries cannot adopt a robotic inspection without a defensible comparison against the human process it supplements, and assembling that evidence is slow, expensive and precisely the sort of thing they publish work packages for.
Capabilities
- Confined-space inspection
- Compliant manipulation
- Evidence packs for regulated industries
- Perception with calibrated uncertainty
- Custom end-effector design
- Recovery behaviour for inaccessible environments
Ways in
The routes that exist, and what each is suited to.
Published work packages
Trials, comparative studies and perception development put to the network, generally where a host site or an accredited laboratory is required.
Trial hosting agreements
Arrangements with maintenance organisations and manufacturers who can provide a real structure with real defects.
Joint certification evidence programmes
Longer collaborations with an operator and a university to build the comparative evidence a regulator will accept.
How to approach them
They want defect sets that are not stock photographs. A partner who can provide genuinely service-damaged structure, with known and unknown defects, is more valuable to them than a funding contribution.
They expect comparative studies to be run honestly, including the cases where the human inspector wins. Partners uncomfortable with that should say so at the outset.
Known for
- A confined-space platform now in supervised use at two maintenance organisations
- A comparative inspection study against human inspectors, published to the participants
- Force-limited handling developed for high-value optical components