Assessing Animal Emotion and Behavior Using Mobile Sensors and Affective Computing

Introduction

Figure 1.

Intensive animal farming

© 2017, SmartStock. All Rights Reserved. Used with Permission.

Affective computing is a young field that large industry is starting to take notice of. For example, Facebook has recently unveiled a suicide prevention application that relies on big data and machine learning to detect the potential for suicide (Constine, 2017). Facebook goes beyond traditional sentiment analysis in examining content posted by users by trying to infer a potential emotional state that might be dangerous to its users. In addition, Microsoft has developed powerful software that can detect emotions based on facial expressions (Parkinson, 2015). The potential for human centric applications is showing a lot of promise. In the context of affective computing and animal sensors, it is not much of a leap to consider the potential of an animal centric affective computing. Animal agriculture in the United States, specifically livestock and poultry, is a major industry in the United States with past and projected annual exports often exceeding $25 billion dollars every year (Cooke, Jiang, & Heerman, 2017). At the very least, an animal agricultural approach to affective computing might be an area of fast-untapped potential. The authors believe an explicit approach to study core affect shared by humans and animals carries advantages that will be discussed later. In short, the authors believe there is rich merit to pursue affective computing in animal sensors.

One of the challenges in writing on an interdisciplinary topic is addressing properly the broad collection of themes. In our case, the specific focus will be on affective computing, animal sensors, human, and animal affect. Therefore, the authors wish to caveat our approach by saying that this chapter is not attempting to be exhaustive in our coverage of ideas, but painting a broad and clear picture. The authors believe that affective computing and animal sensors are still very young and nascent fields. As such, the authors believe the time has come for the explicit and deliberate pursuit of animal affect via affective computing based on animal sensors. In fact, much of the groundwork already exists and the authors desire to show that an affective computing approach to animal sensors is viable and requires an explicit conceptual reevaluation of what animal sensors can do by making affect a core research goal. This chapter shall first proceed by discussing the historical development and contemporary state of affective computing. Second, it shall briefly give similar coverage to animal sensors. Third, the chapter will put affective computing and animal sensors into a broader context by discussing human and animal affect. The chapter will tie these together by discussing an affective computing approach to animal sensors and a potential design for affect and arousal detection using animal sensors.

Affective Computing

Figure 2.

Empatica E4 Sensor

In this section, the authors will provide a brief discussion of affective computing as motivation for its implications in animal sensor research. This chapter shall proceed in a chronological order outlining the intellectual origins of affective computing. As this chapter progresses, it shall outline some of the ever present challenges and future promises of the field. Finally, this chapter will briefly touch upon some potential that is still unexplored in animal sensor research, which will be explored in detail in the last section of our chapter.