Modalities Research other than Text,Image or Voice.

Currently most of the AI works revolves around 3 different modalities. Text - Majority portion Images - Significant portion Voice - Less . Are there any ways we can capitalize on other senses or ways to incorporate other senses in AI : Touch Taste Smell What if we go from some text classifier classifying oranges as rotten or good based on image classification to a multimodal classifier which not only uses camera to take photo and classify but also takes smell into consideration to classify a fruit as rotten or fresh?