I have deep concerns and keen interests in autonomous vehicles. This article is based on recent news published by Daimler AG on Level 3 automated driving’s technical requirements.
The above news caught my eye, and because they did not explain enough, I thought to write a detailed article on the subject.
The time you spend in your private vehicle will transform in the future. Private life in cars will change hell a lot because Autonomous driving is happening more and more. Some fully automated cars are already in prototype mode, and soon we will see them running on the roads.
There are 5 Levels of automation, and I briefly visualized these levels in my last article. You can read the article Data Annotation and the five levels of Automation.
Steering and braking – Drive Pilot?
The first query that comes to everyone’s mind when talking about self-driving cars is how driver pilot steering and braking work.
With the LiDAR sensor and redundant systems… DRIVE PILOT also has redundant steering and braking systems and an inessential onboard electrical system. It remains maneuverable even if one of these systems fails, and a safe handover to the driver can be confirmed. A powerful central control unit comes up with the necessary sophisticated software functions for depending on automated driving. Artificial intelligence is used for image development, for instant. As part of the state-of-the-art safety architecture, all algorithms are deliberated twice.
Next-Generation Car Design Will Require Miniaturized Solutions
Who will dominate the race? Furthermore, how might driverless vehicles change society? While the car business’s eventual fate is as yet unwritten, here is the thing that we know: Industry pioneers should dominate the network to convey the V2X (vehicle-to-everything) abilities of completely self-ruling driving guarantees. Here’s an outline of crucial availability prerequisites to get that going.
To fit V2X frameworks in a car engine, manufacturers should create car-grade scaled-down and profoundly integrated availability arrangements. Sensors and actuators are getting more modest as connections increment.
The Automotive Ethernet enables manufacturers to convey both force and information over a solitary pair of wires, either protected or unshielded. TE Connectivity is critical to affirming the freedom and creepage edges important to achieve this target.
Simultaneously, manufacturers should decide and meet pollution and moisture resistances for frameworks. In-vehicle organizations will require the lightest and most minor fatty link congregations and connectors made at this point – which will be put in ever denser designs. TE helps the car business make these compromises and optimize their answers, to make information navigate frameworks quicker and more forcibly effective, yet with less absolute material connector and cabling material.
With the massive use of sensors in Driverless cars, which sensors gonna be a must-use?
This is because without using automotive sensors, the Level-3 of automated cars could not be possible, but the question is which type of sensors are being used and could be used?
There are mainly three types of sensors 1. Radar 2. Lidar, and 3. Ultrasound sensors are combined with several other sensors and cameras to make Level-3 possible.
Most of companies use LiDar sensors as their Go-to sensor, and even Daimler AG shared they are using LiDar sensors in their linked video (the link is given above). They use Lidar for the optimal distance and speed measurement; however, Lidar is also heavily used for object detection. This sensor is a multi-purpose sensor.
Positioning of semi driver-less cars
Positioning is a big question when cars are on the road with minimum attention from the driver. Mostly, Automotive companies use differential GPS systems to keep track of their level-3 automation.
How to teach car traffic signs and road geometry
While Daimler AG is using its digital HD map for traffic signs detection, road geometry, and other related information, here, most Automotive Companies are relying on getting a lot of traffic signs detection and road data from companies who own a large amount of road data. Then they convert this data to their systems as Daimler AG does.
The common question is, what about using Google Maps?
Well, Since the sky is always blue in the land of Google Maps, there may not be enough weather/light variation to make the images useful as training tools. Most of the training data, he said, would be collected by the cars themselves, in the real-world environments they’re tested in.- says Raj Rajkumar, a roboticist and autonomous vehicle authority at Carnegie Mellon University
How does a partially driverless car see its surroundings?
Cameras come in place when a car needs to see its surroundings. For example, Daimler AG uses the read-HD camera to let their cars see emergency vehicles approaching from behind.
Bosch has recently launched one of the best cameras, named MPC3 multi-purpose camera for autonomous driving, and they claim it is — better than the human eye. This camera is AI-enabled technology and could be replaced by any other low-quality camera for German auto-makers.
Increasingly Electrified Cars will require a new approach to safety:
Guaranteeing information trustworthiness inside practical safety applications is obligatory for independent driving since living souls will rely upon it. As information prerequisites develop, cars will increasingly energize. It is fundamental to guarantee safe information transport on the actual layer level along these lines and give the most noteworthy conceivable physical and information security assurances. Car organizations are examining where and when to utilize optical-fiber-based availability to avoid electromagnetic obstruction. Fiber optic correspondences don’t encounter electromagnetic discharges, crosstalk, or sign mutilation. This implies optical correspondence network stations can be put close to high-control electric links and gadgets and give reliable working without safety concerns or impedance stresses.
Designs Must Be Redundant and Real-time
Self-governing driving will increasingly request increasingly more reliable organization-based constructions, requiring repetitive, continuous designs. These designs will arrange superior bunches in utilitarian areas and be associated with utilizing a focal passage in a fast information spine structure. Gathering sensors and actuators will be coordinated progressively.
To summarize, where almost all technical solutions are ongoing, and some are working fine for level-3 automated cars, some grey areas must not forget. We still don’t exactly know. What about car interconnectivity with other driverless cars? And legal frameworks to meet the final approval criteria for level-3 automated vehicles. Almost every second Automotive Company has challenges and massively works on these hidden factors.
I hope you enjoyed the article and welcome any comments or feedback on it. We at Ahdus Technology also indirectly help make Autonomous driving possible by providing Data-annotation and Data-science services to our Automotive Clients across the globe.
We proudly announce that we are the first Pakistani Data-Science Company that provides the highest data-annotation services in combination with quality control under the supervision of our Data Engineers.