The deployment of AI algorithms in public transportation has been expanding. With different scenarios and applications, it makes transportation safer and more efficient. To perform AI applications, Axiomtek provides a complete line of systems to choose from.
In-Vehicle smart surveillance solution
To enhance the safety of bus commuters, in the next generation, the bus can be deemed as an AIoT ecosystem that includes multiple cameras for smart surveillance solutions; as any anomaly happens, the system can notify the driver and the control center via the fleet management system.
Driver behavior directly influences the safety of passengers, infrastructure, and even the whole traffic. It is where AI can be the most cost-effective application that is worth investing in. With a trained model, the AI inference can monitor the driver’s behavior. Once improper behaviors like keeping the eyes closed for too long, using a cell phone, or yawning frequently, a warning will be triggered. Other cameras that monitor the blind spot can alter the drivers' awareness of reckless fast cars or pedestrians by sending out warnings; in addition, the IP cameras for bus passengers can transmit real-time video to the driver and record in cases of unexpected incidents like fighting.
Also, with the widespread use of sensors, they can be deployed at the bus doors to prevent drivers from closing the door while passengers standing too close and causing injuries.
To realize smart surveillance on buses, there are four musts for an embedded system; it should be certified rich I/O, have big storage space and have powerful computing capacity. Axiomtek’s E-Mark-certified UST510-52B-FL with a 9th/8th Gen Intel® Xeon® processor would be an appropriate fit. A powerful CPU plus M.2 Key E and Key B for wireless connectivity enable the vehicle to perform the applications mentioned above.
The UST510-52B-FL features a fanless design and has rich I/O and expansion capability. With tailor-made modules, it can provide up to 16 GbE PoE ports to connect with IP cameras and sensors. The LAN ports can be changed from RJ45 to M12 type for a firm installation. The system incorporates four SATA drives and two mSATA drives for big storage space for in-vehicle video recordings. Particularly, it supports smart ignition to protect the bus battery and record data from an unexpected shutdown. With wide operating temperature from -40°C to +70°C plus the anti-vibration kit, the durability ensures stable operation.
Traffic monitoring for safety and public safety
Overseeing the traffic flow is crucial for efficiency and public safety. Edge computing systems can receive data on speeding, jaywalking, and driving through the red light through IP cameras and wireless connection, and then identify the hot zones via analysis and capture the images by AI for public safety. Or, the system can capture the plate numbers of cars parking by the red lines or not yielding to pedestrians for fining. More than violations, observing traffic can get a deeper look at the car flow for operation adjustment. For instance, at the busiest intersections, the red and green lights can be switched by the AI decision to keep the flow smooth according to the actual situation. In some advanced cities, we can even see a warning sign pop up at the roadside when a car is too close to pedestrians. Furthermore, to heighten energy saving, edge computing can turn off the lights on roads when no traffic and turn them on according to the sensors when vehicles come for safety concerns.
To process the workloads right at the edge, Axiomtek recommends the rugged IP67-rated fanless system that can withstand the harsh outdoor environment, the AIE800-904-FL-DC, with the operating temperature ranging from -30°C to +50°C. On the field site, the system can acquire and process a great amount of information from IP cameras via PoE to realize AI traffic.
The system features the AI purpose-built SoM, NVIDIA® Jetson Xavier™ NX, with 6-core NVIDIA Carmel ARM® CPU and 384-core NVIDIA Volta™ GPU with 48 Tensor Cores, and an onboard 8GB 128-bit LPDDR4x memory, which can heighten the processing capability to 21 TOPS. With the high computing capability, edge computing applications such as AI public safety like driving through a red light, illegal parking, hit and run or abnormal situations; with AI analyzing the real-time traffic flow, automated traffic light switch can be performed.
The system provides one GbE PoE and two USB ports, and 9 to 36 VDC power input. To transmit data to the server or cloud for further analysis, the antenna openings on the system are N Jack type with a waterproof design for LTE connection.
Bus tracking solution
Today bus tracking solution has become more common and has been expanding to more cities; not only does it help commuters save waiting time, but in rural areas, the bus does not run very often so people don’t miss the bus. If one of the buses in the fleet breaks down on route, the control center can dispatch an additional bus.
Deploying an embedded system at each bus stop is a must; the systems can receive the GPS signal of the bus fleet from the central server and present it to the passengers. More than knowing the location and the estimated arrival time of the bus, AI algorithms can even assist with creating and adjusting more efficient schedules for different routes of buses. As a result, as the algorithms take real-time situations into consideration, the passengers will not need to wait for the bus for too long or see the buses of the same routes coming together at once.
When selecting a system for a bus stop, the most important selection criteria are wide operating temperature and stability. The lightweight and cost-effective ICO120-E3350 is extremely durable under any kind of weather with operating temperatures from -40°C to +70°C. Thanks to its small size and DIN-Rail design, it is easy to install. It is equipped with full-size and half-size PCI Express Mini Card slots to be compatible with a 4G LTE or Wi-Fi module to receive the data of bus locations. Then, via an HDMI port, it can display the bus location, estimated time of arrival and route on a screen at the bus stop. To make passengers feel safe at the bus stop, having DIO on the system allows for adding an emergency button at the bus stop. Once the button is pushed, it will activate the alarming warning light and notify the police.