Variable Height Navigation Mode

At any given position in an urban VE, there must be a minimum amount of sensory cues available to assist navigation. This is achieved by varying the height of navigation (flying) according to the available information.

Theoretical Model:

Consequently, areas rich in sensory cues have one, or often more, of these nodes in close proximity keeping the user low over the cityscape whereas in sensory poor areas navigation height increases.

Attractors

• Organised by type of cue
• ranked by the VE designer in terms of importance (prominance, complexity, historic value, size, etc)
• ability to dynamically modify the assigned importance/strength of each type (according to the experiences, familiarity, interests of individual users)
• visual versus aural means of scanning the environment for attractors (raytracing, vs directional sound sources, etc)

Avatar Flying Height Estimation

Visual cues: attractor nodes are inserted on a low level of detail/boundary like representation of the city to achieve near realtime performance. The placing of the attractors, mainly in terms of elevation, is crucial for the calculations described next:

• Each new avatar position is checked against all visible attractor nodes via a simplified raytracing algorithm
• The relative importance of each attractor is considered
• A simple mathematical formula defines the recognisability coefficient of the particular avatar position/direction of viewing combination.

Experimental Implementation:

Due to the complexity of the proposed model, the actual implementation features a collision detection transparent mesh floating above the cityscape.

Peaks of the mesh denote the minimum amount of censory cues, creating an inverted 3D plot of the sensory cues intensity over the 2D plan.

Pros:

• the computing intensive raytracing algorithm is avoided
• finetuning is possible using online controls

Cons:

• viewing direction disregarded
• varying the relative importance of different types of cues not possible
• mesh generated manually