Back

“Connor’s Paradise” was developed as a part of the Game Studio Semester at Windesheim University of Applied Sciences. It was commissioned by a professor from the University of Twente, specializing in 21st-century skills, for an EU-funded research project. The game serves as a negotiation trainer, simulating scenarios where players must resolve resource disputes between three tribes after a storm devastates their island. The project culminated in Winnovation, Windesheim’s IT project fair, where it won the Innovation Prize, highlighting its potential for real-world application.

The game features:

• Two negotiation rounds: one egoistic and one collaborative.
• Voice integration: Player speech was transcribed and processed through an AI backend to generate real-time spoken responses and trade offers.
• A trade proposal algorithm: Analyzes player behavior from previous sessions to propose strategic trades.

As a Game/Back-end Developer, I was responsible for:

• Built a secure API in Laravel for temporary ChatGPT access.
• Designed and implemented the logic for real-time AI-generated trade responses.
• Contributed to the design of game mechanics and user experience.
• Developed parts of the user interface to ensure accessibility.
• Supported team members by resolving blockers and ensuring smooth progress.
• Assisted the team leader with some leadership tasks.

Tech stack

• Game Engine: Unity, using WebGL for browser compatibility and low system requirements.
• Backend: Laravel, for secure and reliable AI communication.
• AI Integration: ChatGPT, powered by a custom API for real-time, context-aware interactions.

Features

• Voice-to-text transcription in the browser.
• Dynamic trade proposals based on player interactions and historical data.
• A structured storyline contrasting selfish and collaborative negotiation styles.
• Real-time AI-generated responses, with customisable personalities for the tribes.

Challenges

• Ensuring real-time responses while maintaining conversational quality.
• Balancing AI-generated trades to reflect realistic player dynamics.
• Aligning the artistic, technical, and research components of the project.
• Understanding and meeting the expectations of the research stakeholders.
• Optimizing the game for WebGL to ensure smooth performance on low-end devices.

Results

The game won the Innovation Prize at the Winnovation fair, judged by industry experts and faculty.

It demonstrated potential as a negotiation training tool for 21st-century skills development.

Key findings and recommendations were passed on to stakeholders for further research and potential future development.

Learnings

• Backend Development: Designed secure API endpoints and optimized Laravel integration with ChatGPT.
• Game Design: Improved understanding of mechanics, UI/UX, and player engagement strategies.
• Team Collaboration: Worked in a multidisciplinary team and supported the team leader in resolving blockers.
• Stakeholder Communication: Learned to document findings clearly and align deliverables with stakeholder expectations.
• Voice Integration: Implemented browser-based voice-to-text functionality for immersive gameplay.