A Resource Portal for Automotive Dealers, Industry Vendors and Advertising Agencies

The Road for Autonomous Vehicles Is Paved With Next Generation Hardware, SaaS and AI

Demand for autonomous vehicles has forced the auto industry to develop next generation hardware, software, (SaaS), and artificial intelligence, (AI), to produce vehicles that meet government safety standards and consumer expectations. In order to satisfy these critical needs Original Equipment Manufacturers, (OEMs), have engaged specialized Tier I vendors to provide solutions that are beyond their in-house capabilities. 

A variety of integrated technologies are needed to operate autonomous vehicles on public roadways. These mission critical elements are being developed by a growing number of competing vendors.  No single vendor has sufficient expertise to provide best in class solutions to satisfy all of the following requirements:

  • Mapping surface roads with sufficient detail for onboard AI powered software to make split second driving decisions
  • Communication platforms to properly position the vehicle
  • Sensors to collect real time data on road surfaces and obstacles in all types of weather with limited visibility
  • Driverless operating systems for steering, braking and acceleration
  • AI capable of making driving decisions that meet or exceed human capabilities

Vendors must modify existing hardware and software coupled with new technologies to enhance previous applications to meet performance criteria for each of these solutions.  Perhaps the most challenging requirement is the need to map all of the surface roads in the U.S. prior to allowing driverless vehicles to operate in the public space.

Real-time 3D maps and navigation-critical data for autonomous vehicles is already being offered by forward thinking vendors tasked with solving this problem.  The next generation of 4-dimensional indexing of city streets promises to deliver even more robust real-time, street-level intelligence.  These systems will depend on a crowdsourced vehicular sensor network that gathers continually updated 3D scene, change detection and analytics data.

Once the road is clear, so to speak, the next challenge is to develop sensors powered by AI software to extract information for mission-critical applications in autonomous vehicles. Image sensing bundled software solutions must account for all types of driving situations from low light, to bad weather such as rain, snow and fog to ensure 100% visualization by analyzing signals originated from an array of vehicle sensors.

The data provided must consider variables that might not appear relevant at first glance.  Excessive vehicle weight, for example, could have a negative impact on road safety and degradation.  To address this concern next generation solutions must be able to estimate vehicle weight to improve on-board systems such as Adaptive Cruise Control and Autonomous Emergency Braking systems.

Another data set that must be collected, analyzed and applied is the road signature. This includes potholes, bumps and other fixed and changeable hazards due to ongoing construction, pavement failures or road debris.  Big data and crowd sourcing technologies must be integrated to create vehicle horizon predictions including road characteristics such as: grades, tires-road grip level, puddles and other hazards.

Additional data that must be collected in real time and processed by responsive onboard AI for driverless vehicles includes:

  • Tire-road grip level to maximize vehicle control system performance
  • Road hazard identification as the vehicle moves along the road surface to allow dynamic mapping of irregularities
  • Curvature of the roadways to determine safe speeds when approaching and exiting curves while considering car passenger comfort
  • Road grades for hilly roads must be identified and mapped as the vehicle travels on them to enable vehicle system predictive control

Various technologies are being considered to offer heightened visual capabilities in restricted visual conditions.  One group of preferred solutions relies on short-wave infra-red (SWIR). These systems are ideal for operating in difficult driving conditions including darkness, rain, mist and dust. SWIR based cameras allow much higher reliability and precision compared with other sensory solutions but are not favored in vehicles because of their high cost.

An alternative to SWIR is High Speed Ground Penetrating Radar, (GPR), technology.  It was first developed to allow military vehicles to stay on previously-mapped routes by matching GPR measurements with maps of subterranean geology. This enabled precise navigation of vehicles, despite unmarked lanes and poor visibility due to obstructions such as sand and dust.

GPR utilizes data sourced from below the ground to keep vehicles in their lane.  As a result, autonomous vehicles are able to navigate on any road without being constrained by changing above-ground dynamics. Autonomous vehicles equipped with GPR would be able to navigate in extreme weather conditions – including snow, rain and fog.

The final solution for sensors applied to autonomous vehicles will likely consist of a combination of technologies.  Adding GPR to an existing SWIR sensor suite would enable autonomous vehicles to see in a whole new dimension. 

Driverless vehicles appear to operate independantly with self-contained on-board systems.  However, they also need two way communications with cloud based data sources that send and receive data to determine their position along with other mission critical information.

There are discussions in the auto industry regarding a promised 5G network to  connect with autonomous vehicles allowing the Internet to provide two way communication and data transfer.  An alternative plan is to rely on existing satellite technology that is dependent on next generation mobile antenna systems currently in development.

Satellites already in place, along with a new “low orbit” satellite system planned in the near future, suggest that affordable, portable, and efficient antennas and terminals offering high-data-rate communications for ground to air communications will have a role in developing autonomous vehicles. Micronized Satellite Terminals using antenna panel technology combined with state-of-the-art control and tracking mechanisms can provide a fully contained satellite system capable of communications with autonomous vehicles.

The final solution will likely depend on using both 5G and satellite since they each have their own strengths and weaknesses.  Regardless of the outcome, additional solutions are being developed to address this expanding need to efficiently transfer real time data and provide enhancements to communication protocols. 

One example is a system that creates a hidden communications path which expands the bandwidth of the carrier channel up to 3x over existing protocols.  More importantly, it accomplishes this without changing the original message and it simultaneously enables security and BUS authentication.  Securing data and communications with autonomous vehicles is a major concern that must be addressed.

Current intra-vehicular communications have limited bandwidth, legacy immutable protocols and severe data overloads, with no provision for security and cyber-safety. They are unable to determine the authenticity of the sending party of safety-critical control messages. As a result, tomorrow’s driverless vehicle may be subject to an adversarial takeover of controls. This newly developed system promises to eliminate those threats.

Vendors from industries other than automotive are also capitalizing on the growing consumer demand for autonomous vehicles.  They are re-purposing systems and technologies originally developed for different applications.  More importantly, they are directing their diverse experiences and considerable resources to solve problems unique to driverless vehicles.  The result mirrors the impact that the aerospace industry had to incentivize the development of new technology required to reach the moon.       

The road for autonomous vehicles is paved with next generation hardware, SaaS and AI.  The real challenge is to select the best in class solutions that can satisfy unrealistic expectations of consumers in this next generation means of transportation.  Fortunately, the cost in capital and resources can be shared over multiple industries amortized over many generations who will benefit from the scientific discoveries brought about by their initiative.

Share this post