My mum’s always complaining that pets are dirty and messy – Guess it’s time for GUPI!
Figure 1 : GUPI
GUPI is the abbreviation for Guinea Pig and it’s an evolution of “Tamagotchi” that so many of us went crazy over. GUPI is an autonomous robotic pet that is able to avoid obstacles, respond to stroking and noises as well detect infrared beams from the ‘carrot’ gadget that comes with it.
All this is possible with a wide range of sensors. It has 4 legs/wheels which enables it to move freely. GUPI's movements are pseudo-random and it’s even able to engage a maze and come out of it! Infrared sensors in its eyes and nose enable its obstacle avoidance capability. There’s also an altitude sensor in the body and pressure sensor which allows it to respond when it’s being stroked. The ‘carrot’ when pressed, sends out infrared beam that GUPI picks up and homes in on.
The human touching behaviours are classified using haptic interfaces which adopts gridded pressure sensitive conductive ink sheets as shown in Figure 2. The results are used as signals for reinforcement learning in controlling the interactions behaviors with humans. To recognize more sensitive patterns of human touch (e.g. a tickle, scratch, etc.), haptic devices are needed to measure touching pressure at a high sampling rate and at high spatial and pressure resolutions.
Figure 2 : Gridded pressure sensitive ink sheets
GUPI makes 30 different sounds and has 4 different moods; namely the baby state, learning state, happy/normal state and the sleeping state. It is highly interactive and it becomes lonely and scared and goes into hiding if neglected.
The most interesting thing is the unique identification codes in the chipsets that allows it to react to other GUPIs in the vicinity! They might eventually sing and dance with each other if they learn to accept each other and share their food – the carrot which also doubles up as a battery reloader!
It only costs about £35 Cool!
Links : More Info On GUPI
References : Recognizing Human Touching Behaviors using a Haptic Interface for a Pet-robot , Futoshi Naya Junji Yamato Kazuhiko Shinozawa
Alvin Ong Jun Jie