You are hereFamiliar Area Navigation

Familiar Area Navigation


Homing pigeons have long been paradigmatic in the study of avian navigation, but much of this research has been concerned with the mechanisms that allow pigeons to return to their home loft from distant, unfamiliar sites. In contrast, our group is interested in perceptual and cognitive aspects of navigation within areas that birds are highly familiar with. What aspects of the environment do they attend to, how do they perceive and memorise relevant cues, and how are various sources of sensory information integrated into their representation of familiar space? Using our dedicated research colony of homing pigeons and a variety of field experimental techniques, we have been investigating these questions for almost two decades.

Until fairly recently, received wisdom suggested that homing pigeons do not rely on the visual landscape for navigation. Our research challenged this view, first by demonstrating that giving pigeons the opportunity to preview familiar release sites enhanced their homing performance, then later by showing - through precision miniature GPS tracking - that birds' route choices were heavily influenced by the underlying landscape. (See image above, showing a bird that takes a very indirect route home, taking in roads and even going round a roundabout - note that it does this the right way according the UK highway code!)





Our work employing miniature GPS trackers has also demonstrated that through increasing experience with release sites, pigeons come to adopt their own individually-preferred, highly stereotyped routes which they continue to recapitulate extremely faithfully during all subsequent flights. These routes are most likely based on memorised chains of visual landmarks, and the fact that they can differ greatly between birds (see image above right: three tracks from each of four different birds shown in different colours) suggests that route choice in pigeons is not restricted to a single "optimal" route, but instead reflects idiosyncratic landmark choice and learning histories. The structure of these routes, the stages through which they are learned and honed, and the degree of accuracy with which they are recapitulated once acquired tell us much about the way these birds perceive the landscape, memorise relevant information, and ultimately, how they represent space.

In close collaboration with computational engineers we are also active in the development of novel ways of analysing animal movements and the relationship between landscape features and route structure. Employing techniques from machine learning, signal processing and pattern recognition we are beginning to explore within- and between-route variation, and are developing predictive models of routes within the familiar area.