Prf Clara Mancini

Project aims: Following various communications with Carsten Hertwig (Director Sustainable Sanctuaries, CEO BÄRENWALD Müritz gGmbH and CEO Tierart gGmbH) and Florian Eiserlo (Sanctuary Manager TIERART gGmbH) the aims for this initiative are provisionally summarised as follows (to be agreed / redefined with Tierart / Four Paws): 1)To increase the tiger holding capacity at Tierart 2)To advance standards of tiger care and welfare beyond current expectations 3)To effectively utilise the existing physical assets at Tierart

Grounded in prior research and its promising early findings, this project aims to support the development of a canine bio-detection system for human disease diagnostics, through the funding of a PhD studentship in Artificial Intelligence (AI) applied to Animal-Computer Interaction (ACI). Bio-detection with dogs has significant potential for early disease diagnostics but it also still presents significant limitations. This doctoral project will leverage animal-centred interaction to unlock the potential of bio-detection with dogs, bringing together sensor data sources and machine learning methods to detect, interpret and model the nuances of dogs’ olfactory stimulus response to biological compounds. This will lead to dramatic improvements in the early, non-invasive, safe, fast and inexpensive diagnosis of medical conditions including aggressive cancers, bacteria and viruses, by expediting the development of effective bio-electronic sensors for large-scale diagnostics and by informing strategies for the treatment of aggressive cancers, with a particular focus on colorectal cancer as a case study.

AMS record created to record the income for the Seventh International Conference on Animal-Computer interaction ACI2020

Cat Royale is a project to explore how we can design domestic robots to enrich the lives of companion animals through play. The artists Blast Theory and Computer Scientists from the University are creating a bespoke environment to be inhabited by a small family of three cats for up to six hours a day for twelve days. At the centre of this environment will be a small robot arm that attempts to enrich the lives of the cats by playing with them. Their interactions will be filmed, and edited material will be shown to the public as a way of promoting understanding of how future technologies can be designed to deliver positive experiences for companion animals. Our aim is to create an experience that will engage the cats and viewers in a positive way.

The Open University’s Animal Computer Interaction Laboratory is undertaking research to improve the work and experience of Mobility Assistance Dogs. The Dog-Smart Homes Project aims to develop a range of portable canine-friendly interfaces to improve the welfare, training and performance of dogs whose lives are dedicated to helping vulnerable humans. The project will deliver a production-ready set of controls for canine use that can be purchased and retrofitted to make the domestic environment more accessible to mobility assistance dogs and, in turn, to their assisted humans. The project will also advance understanding of the interaction between mobility assistance dogs and their working environments and how such environments could meet the interaction requirements of both assistance dogs and assisted humans. The project will further develop Interaction Design principles and frameworks specifically intended to inform the design of interactive systems intended for canine users. The proposed project is motivated by real life challenges faced on a daily basis by working dogs, who are required to interact with environments that are designed for humans and thus are hardly accessible to dogs. To address these challenges, the Animal Computer Interaction Laboratory (ACI), a pioneer in this area of research and application, has teamed up with UK charity Dogs for Good and has already developed early prototype interfaces that have been tested with positive results. With the partnership and sponsorship of Pet Plan Charitable Trust we now wish to build on our early achievements to deliver a complete toolkit that can be marketed worldwide, to support the daily activities and change for the better the lives of thousands of Mobility Assistance Dogs and their humans.

Cancer detection with dogs, pioneered worldwide by UK charity Medical Detection Dogs, has been shown to have significant potential for the non-invasive secondary screening of cancer in humans, particularly for those cancers that are difficult to diagnose. Cancer detection dogs are trained to recognize the odour of volatile organic compounds from cancer cells (e.g. bladder cancer, prostate cancer) in biological samples (e.g. urine, sweat, breath), and to signal back
to their trainers when they identify the odour marker they are trained to recognize. To communicate with their trainers unambiguously, the dogs are taught to
use signaling conventions (e.g. sitting down in front of positive samples) whose arbitrariness limits the signal’s expressivity and reliability. The dogs need to ‘translate’ their spontaneous response to an olfactory stimulus into an operant response that is alien to signaling behaviours they have evolved as a species. Additionally, the conventions only afford the dogs the ‘utterance’ of binary messages (i.e. “cancer is present” or “cancer is not present”), with no provisions for expressing nuances in between; however, this may not be sufficient to describe the samples. When the dogs’ behaviour deviates from the signaling conventions, trainers are faced with the problem of interpreting the meaning of such deviations. To address this problem, the OU’s Animal-Computer Interaction Lab, in collaboration with Medical Detection Dogs, has developed a proof-of-concept canine-centred interface that uses sensors to capture the interaction of the dogs with the samples. Recorded data indicates a correlation between the samples’ content and the dogs’ response to them; it also indicates that our approach can enable cancer detection dogs to express their response in more spontaneous and nuanced ways, while reducing human interpretation biases, thus increasing the accuracy and reliability of their detection process.

Clara to host a two-day workshop in her ACI labs, discussing different projects including with cancer detection dogs, assistance dogs that operate cash machines and washing machines etc. Focus on how to design with, rather than make animal simply fit into human ways of doing things.

Berkhamsted-based Retrieva has developed a dog tracking collar which has proved very successful for domestic and professional use (search and rescue, policing, etc.). Now we wish to explore the development of a complementary harness which can monitor some of the dog’s vital signs. We envisage creating a “pod” of sensors to be appropriately fitted to a harness, over the chest. As a starting point, we wish to incorporate sensors that monitor respiration, body temperature, body movement, and ambient temperature. The pod would also include a small RF or blue-tooth transceiver that would communicate the data to the collar. The collar would then use its primary means of communication, GSM or its RF beacon to provide vital sign measurements to the dog’s controller or owner. There would need to be some firmware changes to accommodate this, the cost of which would be outside of the project – covered by Retrieva. The contexts in which this could be used range from the general (where a pet owner may wish to monitor the dog when exercising or recovering from illness, or a racing dog trainer may wish to apply fitness science to the training process) to the very specific (where dogs may need to be monitored remotely by police, search and rescue or the military in conflict zones, or where they need to be left in a vehicle or kennel for any length of time). The desired outcome of this project is the investigation, specification and creation of an independently powered sensor pod to measure vital signs that communicates the data to the collar using short range RF or Bluetooth, which can be incorporated into a harness. In the long term, this project could also lead to the development of a more sophisticated wearable device to monitor a variety of vital signs and environmental conditions for a number of uses. This proposal results from the findings of a dog-tracking study conducted by Dr Mancini and examining the use of our technology. The study appears in Mancini et al.’s ACM Ubicomp’12 paper “Exploring Interspecies Sensemaking: Dog Tracking Semiotics and Multispecies Ethnography”, of which we are co-authors. Hence the aims of the proposed work are directly in line with the research interests and portfolio of Dr Mancini, who will be leading the OU effort, and will result in the production of new research papers as well as a likely long-term partnership and future collaborations between Retrieva Ltd and the OU.

None