Application Highlight: ADLINK ADM-TJ30 Brings AI to ADAS

April 08, 2024

Sponsored Story

Despite the fact that most of the hype around advanced driver assistance systems, or ADAS, is focused on getting closer to consumer autonomous driving, the real mass of implementations and driving force behind the technological advancements that have been made recently is in the fleet and commercial spaces.

Operators of large commercial vehicles, public transportation systems, and fleets of cars and trucks are seeking solutions that will allow safer operation, better tracing and monitoring, and deeper data insights that will lead to more efficient processes and cost-saving outcomes.

Modern ADAS is quite sophisticated, of course, allowing for lane departure sensing, automatic collision detection and prevention, and even some level of autonomous driving via assisted steering and cruise controls. What is needed now is a leap forward that leverages next-gen AI processing and applications to proactively make decisions based upon incoming sensor data to choose which ADAS function is best in any given situation, both to make operation safer for drivers, occupants, and surrounding environs, but also to choose more efficient routes, braking, and speeds. An AI-equipped ADAS can ease this burden on drivers, and even take it entirely, leading to improved outcomes.

Application Use Case

Drivers of large commercial vehicles like buses, freight trucks, and heavy-duty construction vehicles must often make many small corrections and maneuvers at speed under high-intensity conditions. These can include inclement weather, congested city driving, highway speeds, poor road conditions, and inattentive pedestrians, among others. An AI-equipped driver assistance application can choose between ADAS functions depending on vehicle speed, road conditions and even input from radar or LIDAR sensors.

As an example, a freight truck in the US in most cases must be operating at no more than 80,000 pounds gross vehicle weight, with no more than 20,000 pounds on a single axle, and 34,000 pounds on a tandem axle group. That means that a fully loaded truck needs more than 500 feet to stop under ideal conditions. With such a long stopping distance, drivers are constantly managing speed and gear position, which is physically and mentally exhausting. That’s true over any distance, but especially near the end of an 8 or 10 hour shift behind the wheel.

Instead, imagine if the driver could instead take advantage of an AI-ADAS that could manage speed based upon road conditions, traffic levels, other drivers’ behavior on the road ahead and behind, and still offer the operator the ability to change inputs and parameters when required. Meanwhile, the central office can have visibility into the vehicle telemetry in real-time or in aggregate while the AI aids in monitoring and alerting for incidents, maintenance issues, or safety concerns.

To accomplish all this, the vehicle would need a powerful and resilient embedded computing device that has the processing power to support the AI-ADAS, and the required near-real-time low latency to make the system safe under any operating conditions.    

ADLINK ADM-TJ30

The ADLINK ADM-TJ30 AI-ADAS solution is designed to accomplish all those objectives while simplifying implementation.

It is equipped with an automotive-grade ECU system, 8 cameras, and a built-in vision-AI deep learning algorithm. The two front-view cameras allow it to support FCW, LDW, and ADAS functions at up to 100 meters, without blind spots. The four sideview cameras bring in two sides of BSIS (UN R151) to protect other road users, especially pedestrians, bicyclists, and motorcycles. The rear view camera adds visibility at the rear and combines with the rest of the vision network to support AVM. Finally, the in-cabin IR camera monitors the driver for more than 10 kinds of driving status conditions and can recognize up to 100 different drivers to prevent unauthorized access or operation.

All that sensing capability adds up to an all in one system for 360-degree external and in-cabin intelligence in one ADAS solution and ECU. ADLINK has made it support multiple ADAS functions, including AVM, BSIS/BSD, DMS, LDW, and FCW so there’s no need for companies to deal with multiple vendors for installation and maintenance. It complies with UN regulations R130, R131, R151 for large commercial vehicles that cover lane departure warnings, forward collision warnings, and the blind spot information system. Conforming with such regulations is of course critical in ADAS-equipped large commercial vehicles.

The ADLINK ADM-TJ30 full specifications include:

SoC	TI TDA4V MidEco
Display	1x Display port
CAN bus	4x CAN/CAN-FD
Ethernet	1x Gigabit (RJ45), 1x 100 Base-T1
GPIO	3
USB	1x 3.0 Type A
Storage extension	Micro-SD card slot
Operation Voltage	DC 9-32V
Power Consumption	15W
IP rating	ECU: IP51, Camera: IP67/69K (DMS camera: IP5X)
Operating Temperature	-40℃ ~ +85℃
Storage temperature Range	-40℃ ~ +95℃
Reliability	ISO 16750-4 / IEC 60068-2
EMC	ISO 16750-2 / ISO 10605 / VSCC 56-3

Additional Resources:

• Product Page:  https://www.adlinktech.com/Products/Automotive-Computing/AI-ADAS/ADM-TJ30?lang=en
• Data Sheet: https://www.adlinktech.com/Products/Download.ashx?type=MDownload&isDatasheet=yes&file=2165%5cDatasheet_ADM-TJ30_20240325.pdf
• Solution Brief: https://www.adlinktech.com/Products/DownloadPublication?lang=en&ID=2636&pdNO=2165