Summer Creative Practice Course
/
Personal Project
Project Context
This project is a personal exploration inspired by my dog Better, whose hind leg weakness made it difficult for her to walk, requiring her to relieve herself indoors. I finished the project during the summer creative practice course.
Output
Data Collection Units, Database, Data Learning Model, Data visualization, Mechanism Prototypes, User Testing, Product Modeling and Rendering.
May.2024 -Sep.2024
Beijing.China
What Problem Am I Solving?
Training Dogs to Relieve Themselves in Specific Spots Indoors
Helping dogs learn to urinate in specific spots at home independently through the designed mechanism, reinforcing behavior without direct owner supervision.
Why Am I Doing This?
To assist the dog and owner with managing indoor toileting.
I have a 14-year-old dog named Better. As she has aged, her legs have weakened, making it difficult for her to walk. I wanted to design a device to assist her with indoor toileting and provide convenience for my family.
Training dogs to use specific spots indoors is essential for both the dog and the owner. The goal is to create a low-cost, easy-to-learn system that dogs can use independently when the owner is unavailable.
User Research —— 57 Questionnaire Responses from Dog-owning Households
01 Basic Information about Dogs
46.43% are medium-sized dogs. 42.86% are aged 1–3 years. 66.07% live in apartments.
02 Average Walking Frequency and Duration
80.36% spend less than 30 minutes on each walk.
75% of respondents walk their dogs fewer than three times a day.
03 Training Status and Reasons for Not Attempting Training
60.71% attempted to train their dogs but failed. 25% successfully trained their dogs.
37.5% gave up due to a lack of time or energy. 12.5% were unsure how to start training.
12.5% felt no need for training as they could walk their dogs three times a day consistently.
04 Challenges and Reasons for Training Failures
71.43% faced conflicts between training and other commitments, such as work, making it difficult to continue.
72.92% mentioned that working during the day prevented timely feedback, leading to poor training outcomes.
78.57% reported frequent accidents during training, indicating their dogs struggled to adapt.
60.42% had to frequently clean up after their dogs' inappropriate toileting, adding to their burden.
05 Methods for Successful Training and Desired Features in the Device
75% use reward mechanisms (e.g., treats or toys).
100% incorporate attractants (e.g., urination inducers).
53.57% want the device to include behavior monitoring and reminders, such as tracking success in toileting.
Design Concept — A Guided Toileting System with Behavior Reinforcement
Three Parts of the Whole System:
Data Collection
Collects data on the dog's water bowl and behavior, then sends it to the server.
Data Learning Model
Analyzes and predicts urination behavior, sprays pet urination attractant.
3. Reinforcement Prototypes
Dog urinates in the correct spot, reward mechanism activates, like treat dispensing.
Part 01 — Data Collection
Gathering Dog Data to Predict Toilet Habits
Unit 1: Dog's Motion Collection Unit
Tracks the dog’s movement and sends the data to the server.
Unit 2: Water Level Collection Unit
Monitors the dog’s water bowl and sends the data to the server.
Code Snippet for Data Collection
Actual Data Collection Process
Database for Storing Collected Data
The ESP8266 board collects data and sends it to the server, where it is stored in an SQLite3 database.
Part 02 — Data Learning model
Outputting Predicted Results to Control the Water Pump and Spray Urine Attractant
Data Parsing
Processes sensor data for model training.
Model Structure
Built using Paddle, Baidu’s AI framework.
Data Learning Visualization: Visualizing Urination Data and Patterns
Data Model Prediction Accuracy
The data model's success rate is around 75%, given the limited data and the complexity of the dataset.
Part 03 — Reward Prototype for Positive Reinforcement
Unit 1: Receiving Results from the Data Model to Control the Water Pump and Spray Pet Urine Attractant.
Unit 2: Providing Positive Reinforcement by Dispensing Treats and Playing Recorded Encouragement.
When the sensors detect that the dog has urinated in the correct spot, the servo will rotate to drop treats, and the MP3 player will play the owner's voice as encouragement.
