In the last decade, the automobile’s navigation system or in-vehicle infotainment (IVI) has become more popular in the market. Recently, even entry-level vehicles are installing an IVI system with large displays. We are seeing many automobile manufacturers starting to emphasize their latest IVI system features: 10 inch display, premium sound system, touch screen, connect to Google map, voice recognition, and so on.
The latest IVI systems are more likely to focus on two elements which are important differentiators to increasing the sex appeal of new technologies in the vehicle. One is delivering new Human Machine Interface (HMI) functions, and the other is connectivity. In order for new vehicles to be attractive products for the younger generation buyers who happen to be the very tech savvy, automobile manufacturers need to provide the same user experience that is provided by the consumer markets in mobile and tablet devices.
When thinking cutting edge IVI, Tesla comes to mind for most people, they introduced a 17" capacitive touchscreen that features various media content and Google map apps. Mercedes also introduced a noteworthy IVI system that has two 12.3”capacitive touchscreens - one panel is for virtual gauges and the other panel for navigation, media and HVAC control. Such luxury vehicles reveal the future IVI trends in advance which are later adopted by the mid-level and higher production volume vehicles.
The recent adoption acceleration of such systems also brings new concerns. A bigger display may become more distracting over dedicated push-button control. Simple operation is important to reduce driver distraction. The U.S. Department of Transportation (DOT) and National Highway Traffic Safety Administration (NHTSA) are concerned about diver distraction. An increase in display sizes have also increased concerns of driver distraction due to complicated multitask operation during driving. In 2012, both DOT and NHTSA have unveiled voluntary guidelines for car OEMs, called the “2-second rule”.
These guidelines are aimed at encouraging IVI systems to be designed so tasks can be completed by the driver while driving with glances away from the roadway of two seconds or less, and a cumulative time spent glancing away from the roadway of 12 seconds or less. (refer to Docket No. NHTSA-2010-0053). This guidance is just the first steps by the DOT to help reducing distracted driving. Future IVI systems must have more robust user-friendly HMI for drivers, so they can recognize all safety-critical information easily and operate multitasks without distraction.
Voice recognition (VR) would be a key technology to help achieve these guidelines. Voice recognition can allow a driver to access many command and control features such as the radio and HVAC systems without taking his/her hands off the steering wheel or eyes off the road. This also means VR technology needs to support natural conversation, multiple accents, and multiple languages with high quality. This is not a trivial task otherwise it would exist in cars today.
Additionally, there is the potential of combining VR with gesture technologies, which could add yet another dimension to advance user interfaces. If deregulation of the windshield moves forward, it would allow the windshield to be used as a big screen to display a lot of information - this is more commonly referred to as Heads-Up Display (HUD). If you can remember in the movie Mission: Impossible 4, you could see an example of graphical information projected on the windshield, but instead of straining to touch the windshield to issue commands, would it not have been cooler to utilize gesture control? It makes sense for automotive manufacturers to develop new technologies such as VR, gesture and HUD. These need more computing power so GPU computing will be needed. Using GPU computing correctly can perform more complicated algorithms more efficiently than a CPU-only implementation.
The second key differentiator in the vehicle is its connected car capability. Audi and Qualcomm have announced in-car 4G LTE wireless broadband that will deliver internet services, that includes access to latest information such as maps - often maps based on DVD or other media formats can become outdated compared to maps provided by internet services. Next generation connected cars will recognize each other, through Vehicle-to-Vehicle (V2V) communication technology. DOT is very positive and supports V2V technology for light vehicles to help save lives, prevent injuries, ease traffic congestion, and improves the environment. It requires six in-vehicle components of Dedicated Short-Range Communication (DSRC), GPS receiver, memory, driver-vehicle interface, vehicle’s internal communications network and safety application ECU.
ARM provides solutions with scalable CPUs and GPUs to support the demands of IVI and the connected car. The Cortex-A series, which is used for most smartphones and tablets, can provide enough performance with lower power consumption. Indeed, the Cortex-A series has more than 85% market share in SoC of IVI - and more, the latest ARMv8-A supports 32/64-bit application architectures which can achieve even higher performance. We have launched the Cortex-A57 processor for high-end and Cortex-A53 processor for power efficiency. ARM has a GPU called Mali, which is commonly used for Android devices and the ARM Mali-T700 GPU series supports GPU computing.
A baseband processor is needed for communication and that can be covered by the Cortex-R series which are widely used for mobile phones. To support high-speed LTE or 4G and multi-channels of communication of V2V requires the CPU to be high performance, low-power and functional safety. ARM is focusing on multi-operating systems that will ultimately run on the same platform. The Cortex-R series are used for ABS or other safety critical chassis applications – but the recently unveiled ARMv8-R architecture supports 32-bit architecture and will meet the new automotive trends to support hard real-time, functional safety, and hardware virtualization. We believe ARMv8-R will be essential for the connected car, and of course ARM is also targeting other applications.
The new era of IVI will come by 2020 with new HMI and V2V services in vehicles. Cutting edge voice recognition will enable natural conversation and easy operation - road conditions, vehicle conditions and maps will be displayed on a windshield controlled by gesture motion. ARM’s CPUs and GPUs are ready to support the future.
Vestec and ARM brings Natural Conversation and Dialog VR to IVI designs in vehicles.
I have to agree with Soshun’s blog in reference to VR (Voice Recognition). However, VR needs to take a leap forward in design. What is needed is that the user can speak to the application as if it were another person. The systems need:
Vestec’s solutions to VR solve these issues leveraging state of the art advances in Artificial Intelligence (AI) and Machine Learning (ML). Vestec’s solution extracts important keywords and identifies the context in which they are spoken. Vestec’s technology recognizes new contextual associations between words previously encountered and retains them for future use.
Vestec has created a unique enhanced filler model which is designed to highlight keywords recognition while absorbing and suppressing extraneous distractions such as irrelevant words and noise.
Safety is the key factor here so if we can create a user experience in which natural voice interface can directly and easily create actions and results we will have less distracted drivers.