Murda.ai

We have developed a 3-D, single player, mystery-puzzle game using the Unity engine. The game-play consists of the player who has just committed a murder, at the scene of the crime along with 3 NPC bystanders, whom are all being interrogated by a Detective NPC. The objective of the game is the for the player (criminal) to successfully convince the Detective of his/her innocence, and essentially lie their way to freedom. A "suspicion meter" has been implemented to measure the player's level of suspicion from the Detective's perspective. By default, this meter will begin at 50%, and must reach 0% for the player to be considered "innocent". The player will be aware of the scenario through looking at the in-game mini-map, and also listening to the dialogue between the 3 innocent NPCs and the Detective. When answering the Detective's questions, the player must avoid answering with lies that contradict his/her own story based on the information learned from the surrounding environment. The Google Cloud Speech-to-Text API is used to detect the player's speech (through their microphone) and convert it into text. This text version is subsequently passed to the Google Cloud Natural Language and IBM Watson Tone Analyzer APIs, which will decipher what the user said and output quantified metrics (i.e. measure of emotions, sentiment, etc.). Based on these values and their relation to the dialogue between all characters, a back-end algorithm will determine how much to increase or decrease the player's suspicion meter. Ultimately, the player wins and escapes when his/her suspicion meter reaches 0%, or loses and is detained by the Detective if his/her suspicion meter reaches 100%.

Built With

• google-web-speech-api
• ibm-watson
• javascript
• python
• unity