Description:
Reference #: 1666
The University of South Carolina is offering licensing opportunities for All Weather Traffic Monitoring with Roadside Infrastructure.
Background:
In 2021, a staggering 7,500 pedestrian fatalities were reported in the United States as a result of vehicular collisions. According to the US Department of Transportation, over 50% of fatal or injurious road accidents occur at or in close proximity to traffic intersections. Most, if not all, of these deaths and injuries can be prevented by proactively warning the drivers, vehicles, and pedestrians; for example, by notifying the pedestrian of oncoming vehicles during cross walk or by enabling smarter speed control for vehicles near traffic interactions. While the advent of full driving automation (i.e., Level 5 autonomy) holds promise for a future without such tragedies, there is a pressing need for an interim solution at intersections to reduce the frequency of these incidents. Such a system can also collect important statistics and telemetry information, such as real-time pedestrian and vehicular traffic at intersections, their speeds, vehicle proximity to intersection stop bars, occupied lanes, and vehicle types, which can enable a variety of applications related to traffic monitoring and management. Existing vision-based sensors, such as cameras and LiDARs, provide powerful tools to not only measure such traffic behavior at intersections but also improve pedestrian safety. But the performance of the vision-based sensors are often significantly impaired by the scene conditions, such as no ambient lights or poor visibility during night time, heavy rain, or dense fog.
Wireless signal based object detection systems can alleviate such a problem. A wireless device can illuminate the target scene by transmitting wireless signals and receiving them bounce off of different objects. Based on the time-of-flight and angle of reflections, this device can map the entire environment and “see” the static and dynamic objects within it, even under low visibility and poor weather conditions. Next-generation wireless networking devices operating at higher frequency, such as 5G picocells, offers a solution to this issue. These networking devices have built-in millimeter-wave (mmWave) technology, which offers a substantially higher data rate than traditional wireless technology and can host multiple, palm-sized antenna arrays to create hundreds of beams for serving mobile users. Due to the short wavelength and wide bandwidth operation of mmWave signals, each picocell can also function as a high-precision environment sensor. So, these devices can be augmented into roadside infrastructures, particularly at traffic intersections, to provide high resolution monitoring of vehicles and pedestrians. MmWave devices provide an advantage over camera-based systems during poor weather and low visibility conditions, as wireless signals can penetrate through some obstructions like dense fog, while lights cannot. So, the ubiquity of mmWave technology in 5G-and-beyond devices, such as the picocells in roadside infrastructure, enables the opportunity to bring traffic monitoring and pedestrian safety at intersections in all weather conditions. However, the design of mmWave sensing on networking devices presents two challenges.
First, although mmWave devices are good environmental sensors, it is difficult to simultaneously run sensing applications and data transfer. For instance, if a pedestrian walks in front of a mmWave picocell while it is streaming data, it can disrupt the Line-of-Sight (LOS) communication path. While its beam can be steered towards the Non-Line-of-Sight (NLOS) path or networking and sensing operations can be time-multiplexed to reduce interference, these can negatively impact both pedestrian detection accuracy and network performance by reducing throughput, increasing latency, and disrupting the delivery of packets to critical applications. A strawman approach for networking-sensing coexistence is to augment devices with special-purpose sensing hardware to use different parts of the mmWave spectrum and avoid interference. But this will prohibit deployment of the sensing applications to a large number of existing and future inexpensive mmWave devices.
Second, mmWave devices are vulnerable to more specular and variable reflectivity challenges (compared to Wi-Fi or LTE) due to their high-frequency operations. So, depending on the location, orientation, and absorption properties of objects and pedestrians on the road, the signals transmitted may not reach back to the device. This can result in a loss of information about objects and pedestrians, as well as difficulties in accurately capturing their properties.
Invention Description:
This invention is a system that enables coexistence of networking and sensing on next-generation millimeter-wave (mmWave) picocells for traffic monitoring and pedestrian safety at intersections in all weather conditions. This system proposes using 5G picocells, which operate at mmWave frequency bands and provide higher data rates and higher sensing resolution than traditional wireless technology. Furthermore, this system customizes deep learning models that not only can recover missing information about the target scene but also enable coexistence of networking and sensing.
Potential Applications:
This system can be installed on roadside infrastructure as an environmental sensor, alleviating possible risks that cause vehicular accidents.
Advantages and Benefits:
This system provides higher data rates and higher sensing resolution than traditional wireless technology. This system also can run sensing applications and data transfer simultaneously, which current systems cannot do.