Thursday, 7 July 2016

Noahs Ark audio installation

Ashley Bryant helping fix the acoustic buttons
Final buttons from inside enclosure
Ashley and I installed the buttons inside Noahs Ark elephant enclosure between 11-1, when the animals are kept outside.

Initially, the system was working, and we were able to film the elephants coming into contact with it for the first time.  However, although it was clear they could use the interface and were naturally drawn to explore the surface, the device did not work consistently.  It stopped, and needed to be reset a few times, but then when triggered, one of the buttons played continuously before stopping (it is supposed to only play when there is contact with the sensor).  I suspect that the large amount of metal in the enclosure is having an effect on the cap sensing.

Diagram for keepers
I left a diagram to show keepers how to change batteries and switch system off if it stopped working properly, although I asked them to keep the Arduino charged so we could continue to use the datalogger.

Thursday, 30 June 2016

Making radio buttons

This month has been all about designing and constructing the elephant radio system that we plan to install at Noah's Ark Zoo.  It's tricky, because there are quite a few considerations to take into account.

There won't be a mains supply, so one of the challenges is designing a system that can be battery powered for a week.  This means I can't use PIR or ultra-sonic sensors, which use up a lot of juice, pinging all the time while they wait for an interrupt.

I plan to do some data-tracking, but the data logger shield ( uses many of the Arduino pins, so I need a system that works on top of this simultaneously.

I opted to go for capacitance sensing again, this time using a shield ( that re-calibrates every time it is reset and also manages some simple filtering of the signal to avoid interference.  I was able to add both shields to Arduino Uno and use six sensors to trigger separate audio files from a simple sound board. (

Prototype hardware system
The system relies on 3 different charging systems - the Arduino uses a USB-5V/1A bank; the sound board needs 3x AA (4.5V) and the amplifier+speakers uses 4x AAA (6V).


Lisa and Ashley wanted to do a test with elephants whereby they are offered control over sounds that play - using the same sounds that were tested in a previous study at Blair Drummond.  In that study, the elephants were played 3 different sounds (whale song, classical music and heartbeat) and their responses were noted.  I asked which classical music track was used, but was not able to clarify this. 

Intuitively, the nature of the music seems to me to be important - eg. Brahms v Wagner - hardly the same kind of noise.  I checked online and found some footage and stories about both zoo-housed and sanctuary elephants apparently listening to and enjoying live musicians playing classical tracks:

Sandra at Noahs Ark had explained that they might need to separate the young bulls, in which case we should have two identical systems so they could each have the opportunity to play.

Lisa said it might be appropriate to play elephant noises to the elephants (although I had been advised earlier not to do this) so I checked for some "playful" sounds that elephants had made in a happy context, finding a selection of low rumble-coos, made by mothers to pacify their offspring.  Then Lisa suggested we should drop heartbeat in favour of a classical track.

In the end, I selected [1] rumble-coo; [2] humpback whale song; [3] short clip from Bach D Min for 2 violins.  

Thursday, 26 May 2016

Playful UX for Elephants

Elephants at Dublin Zoo
This week I went to Dublin for Measuring Behavior 2016, to take part in the ACI symposium, which was jointly organised by Ilyena Hirskyj-Douglas from University of Central Lancashire with Anton Nijholt, Patricia Pons and Adrian D. Cheok.  Ilyena has blogged the event in detail here:

I presented "Playful UX for Elephants":

ABSTRACT: This case study describes approaches to the challenge of designing interfaces for an elephant that enable her to control playful systems in her enclosure, for the purpose of enriching her environment. Our contribution to the symposium will showcase the progress of the enrichment toys and explain in detail how we have collected feedback during participatory design sessions with our play-tester Valli, a female Asian elephant. We have attempted to gain information about her enthusiasm for interacting with different systems and also establish how effectively she can use different interfaces by measuring her responses during the sessions.

While I was in Dublin, I visited the zoo and met Gerry Creighton, the Operations Manager, who showed me round and discussed what kinds of enrichment he envisaged for his herd.

Monday, 16 May 2016

Noah's Ark Elephant Eden

Location for acoustic enrichment system above water troughs
Ashley and I went to Bristol, to visit the Noah's Ark elephant facility and speak to the head keeper about installing some acoustic enrichment over the summer.

It was a productive trip - the set-up for elephants is great, offering opportunities to attach a button-operated system to wire between steel bars inside the shed.  Sandra, the head keeper, was very enthusiastic and helpful.  Ashley will be doing some observational studies from the balcony, where she can attach a video camera.

Janu and Machanga browsing together
There are 2 African males at Noah's Ark - Janu (10) and Machanga (7), from Port Lynne and Sweden.  Because they are still young, they sometimes become aggressive and have to be separated at night.  Ideally, Sandra would like to keep them together and free to roam outside or inside during the day and night.  In case they need to be kept inside and separated at night, we will install 2 systems in the enclosure, one on each side of a dividing fence.  This would mitigate against competition, as they could both access the device.  But it's not a scalable solution.

Thursday, 12 May 2016

Spring showers in Skanda Vale

Shower control buttons on the balcony
This week's visit was focused on making Valli's shower controls work.

I had pre-assembled the textile buttons, fitted with ultra-sonic sensors, but we needed to fix solenoid valves to the water pipes and wire everything up to the micro-controller.

Brother Stefan had put aside some time to help with the plumbing and everything seemed to work... until the water pressure dropped.  The Skanda Vale water supply is pumped from a bore-hole, rather than mains, so the pump only kicks in when the pressure is becoming low.  Unfortunately, the valves are designed to work at full mains pressure only.  We need to replace with latching valves.

Fine mist shower spray
One button activated a jet of water, which hit Valli unexpectedly on the back and was not very popular.  She left the shed and refused to come back in.

The keepers concluded that she would have more interest in controlling a water supply if she was thirsty or already in the middle of a bath, so we tried again after her walk.  The high jet was not trunk-accessible and will be lowered to a visible height.
Valli operates the button

When Valli had been showered by the keepers and was already wet, she appeared to be quite content to touch the other button - activating a fine mist spray.

After I left, Brother Stefan told me that Valli had managed to pull out some of the wires from the Arduino, which was very resourceful of her, since they were on the other side of the balcony rail.  Presumably she could grasp the trailing wires through the wire netting.  Brother Danny is keen to experiment with Arduino, so I'm sending them the sketch (well-commented) + a Fritzing circuit diagram + 2 new latching valves.  Hopefully, they will be able to make some independent progress with showers and I will focus on acoustic and haptic enrichment.

Here's Valli with Br Peter, playing with a stick during our walk.

Sunday, 8 May 2016

HCI goes to the Zoo

Clara Mancini kindly delivered our paper, "Trunk-enabled Toys" at the "HCI goes to the ZOO" symposium at CHI 2016 this week (DOI: 10.1145/2851581.2856485 ) .

ABSTRACT: This work explores the use of technology to help create different kinds of interfaces for controls aimed at captive elephants, enabling them to interact with digitally enhanced playful systems (smart toys). The focus of the paper is on current participatory design sessions with an elephant and her keepers, giving rise to insights on species-specific interfaces and how to enable playful encounters with technology.

Wednesday, 4 May 2016

Twycross Zoo

Elephants at Twycross Zoo
On Wednesday, I visited Twycross Zoo with Ashley Bryant, a vet student working with the EWG (Elephant Welfare Group).

Ashley's research involves testing a prototype behavioural analysis tool that has been developed by the EWG.  She and Lisa would like to test some novel acoustic enrichment that affords elephants control over sounds in their enclosures.

We're planning to work together at some zoos, using one of my acoustic enrichment systems and analysing elephants' reactions using their paper-based tool. 

So far, Twycross and Noah's Ark have expressed interest in taking part.
Discussing browsing holes with keepers for possible location
Twycross have a herd of 5 female elephants including a 2 year old calf, Esha.

We had a discussion with the keepers about nature of a button activated system and the most appropriate location for a novel enrichment device.

I showed them a demo button, similar to ones we might deploy with their elephants, and that was ok.  Then we had a look at the enclosure to find a good place to put them.

The keepers would like to use two browsing holes in the wall of the elephant shed.  There's a metal box behind the holes, which can be accessed easily and provides a place to secure a robust set of buttons. Andy, the Head Elephant Keeper, would like to be able to deactivate the device when they need to focus the elephants on a different task.

Demo button
None of the keepers are keen to spend a lot of time filling out the behavioural tool form, but we could use video cameras tracking data instead.

The problem with the browsing holes is twofold:
[1] Elephants associate them with food, so it would be tricky to assess the effect of putting a non-edible system in that location.  Elephants might just be searching for food when they used it.
[2] That corner of enclosure is a blind spot for the wall-mounted video cameras, so we wouldn't be able to track which elephant was using the device overnight - although we could tell if it was being used.

Lisa Yon, from EWG, is planning to discuss other options with Twycross this week.  Lisa would like the enrichment to offer the animals a choice of 3 sounds - heartbeat, whale song and a classical music track.