We examine how issues of investigative and experimental parity between real-world domain science and virtual reality (VR) involving human-environment behavior might be advanced, particularly in the use case of safety science for road-crossing. Our contribution centers on a VR-based traffic flow simulation to recreate, with high fidelity relative to the real world, dynamics of hyper-local interaction between traffic, people, and the roadside environment. An initial demonstration of the system shows that 22 participants responded with high levels of presence, and with high propensity toward natural behavior across road-crossing dimensions. We report these findings even with low-resolution graphic elements. Our results highlight that high levels of user-identified situational verisimilitude (i.e., appearing authentic, particularly to the senses) can be achieved, even with low-resolution graphical depictions. The key, we argue, is the design of appropriate low-level action models to drive user embodiment relative to VR assets. We contend that this finding has wider relevance to consideration of potential channels for VR experience more generally.

This paper reports on the development of a flexible 5-MBd equalized M-QAM testbed for high-speed wireless data communications. The unit operates in real time and was field tested in typical indoor environments. A total of 3600 independent experiments were conducted using the testbed where one of 4-, 16-, 64-QAM constellations were transmitted, and the performance as a function of adaptive equalization and antenna selection diversity was studied. The experimental results presented here help put previous simulation and analytical results into perspective and demonstrate some of the performance bounds associated with a practical implementation. The field trial results show that with only 10-mW of transmit power, reliable 10-Mbit/s data communication can take place in between rooms with a coverage radius of 17 m, and better than 15% outage at an uncoded bit error rate (BER) of 10/sup -3/. The addition of two-branch antenna selection diversity to the system would allow 10-Mbit/s transmission at better than 3% outage and 20-Mbit/s transmission at 10% outage in the same environment. Moreover, 30-Mbit/s data transmission is feasible when both the transmitter and receiver are located within the same room, albeit at higher outage levels. In general, the results demonstrate the tremendous impact that adaptive equalization can have on the achievable performance of indoor links. Average SNR was improved anywhere from 8 dB to 12 dB depending on the particular environment and data rate. The impact on the outage performance, however, was much more dramatic.

The accuracy with which simulation techniques can be used to predict the performance of an experimental high speed M-QAM testbed in typical indoor wireless environments is studied in this paper. Ray tracing techniques were used to produce impulse responses for specific transmit and receive antenna locations where actual measurements were taken using a highly versatile equalized M-QAM testbed operating at 10 Mbps, 20 Mbps and 30 Mbps. The results obtained using a simulation of the testbed plus the ray-traced channel are compared with those obtained with experimental field trials. The paper presents the results of such comparisons for a total of 2,322 trials, where a trial is defined as a simulation run plus a field test for a given data rate and antenna location. The study shows that when quantization noise is factored into the simulation, the performance predicted using the simulation approach closely matches the experimental data.