Time of day
Author: Kelly McGuire
Q: Why did the pedestrians cross the road?
A: They were hungry.
Do you ever become irritable before lunch? We feel you. Also, if you live and work in the Boston metropolitan area or one of the other cities where we’ve gathered footage, we might have seen you on your way to grab a bite to satisfy your growing hypoglycemic-rage. Here at Perceptive Automata, we have a unique interest in the behavior of pedestrians. While we focus on predicting their behavior for the purpose of successful transportation automation, we also get to take a peek inside of the motivations that drive human behavior using the huge amount of pedestrian data we collect.
Scientists have known for a long time that hunger biases our attention to food-related cues in the environment¹ ² ³. Hunger also simply drives food seeking behavior⁴ ⁵. It is the motivation to get food. Imagining how this focused drive to feed might interact with traffic at lunchtime might be a little scary. Food-seeking zombie-pedestrians would pose a risk to themselves and vehicles if they needed to cross a street to reach their destination. So, what do pedestrians do at lunch?
It probably comes as no surprise that we find many more pedestrians on the street around lunchtime. Take, for example, a dataset of a few hundred thousand pedestrians we have tracked in the city of Boston. Simply by tracking their density over time, we can see that there is a huge spike in the pedestrians we encounter for every minute of driving around noon (and a slow climb towards the end of the work day — but that will have to be saved for another post).
This is cultural — American city dwellers eat meals around noon — but it is also based on very simple biology — if the last meal you ate was in the morning you probably can’t make it much past noon without getting a little hangry. At the risk of over-interpreting our data, we can look a little closer at this trend. Is there a difference in human behavior beyond simply being outside as noon approaches and blood sugar falls? At Perceptive Automata, we build State-of-Mind AI (SOMAI) models to understand human behavior. Of particular interest to us (and anyone else in the autonomous vehicle industry): 1) Do pedestrians have a high intent to cross the street (around lunch hours)? And, 2) Are pedestrians aware of the vehicle in the road? This is critical to understand for the safety of pedestrians and the passengers of autonomous vehicles. The short answer is, yes. We see a higher density of pedestrians with a high intent to cross the street and thankfully, this is accompanied by a higher density of pedestrians that are aware of the vehicle in the road at lunchtime as well. This trend remains even when we control for the number of pedestrians we see at different times throughout the day (e.g., more eager road-crossers could have just been due to more pedestrians in general), even though statistical certainty has to be shelved at this point.
Not surprisingly, people have complex behaviors that fluctuate throughout the day. While lunchtime pedestrians may not be food-seeking-vehicle-ignoring zombies, the same pedestrian standing at an intersection at 11:00 a.m. or 12:30 p.m. might be thinking very different things. It might be overkill to guess what they are going to order for lunch, but understanding when they are likely to cross the road (and the path of an oncoming autonomous vehicle) is critical for the success of the autonomous automotive industry.
References:
[1]: Higgs, S., Rutters, F., Thomas, J.M., Naish, K., and Humphreys, G.W. (2012). Top down modulation of attention to food cues via working memory. Appetite 59, 71–75.
[2]: Cornier, M.A., Von Kaenel, S.S., Bessesen, D.H., and Tregellas, J.R. (2007). Effects of overfeeding on the neuronal response to visual food cues. Am. J. Clin. Nutr. 86, 965–971.
[3]: LaBar, K.S., Gitelman, D.R., Parrish, T.B., Kim, Y.H., Nobre, A.C., and Mesulam, M.M. (2001). Hunger selectively modulates corticolimbic activation to food stimuli in humans. Behav. Neurosci. 115, 493–500.
[4]: Burnett, C. J. et al. Hunger-Driven Motivational State Competition. Neuron 92, 187–201, doi:https://doi.org/10.1016/j.neuron.2016.08.032 (2016).
[5]: Lee, Y. H. et al. Food Craving, Seeking, and Consumption Behaviors: Conceptual Phases and Assessment Methods Used in Animal and Human Studies. J Obes Metab Syndr 28, 148–157, doi:10.7570/jomes.2019.28.3.148 (2019).
