‘Right-Sized’ graphics for automotive solutions
Defined value in automotive instrument panels
As part of digital instrumentation, elements such as switches, buttons and analogue indicators are being supplemented or replaced by displays, with traditional HMIs (human/machine interfaces) becoming GUIs (graphical user interfaces). This display, which also has sensory and cognitive capabilities, is the GUI between driver and vehicle. GUIs offer enormous flexibility. The scope and variety of information that can be displayed are almost limitless. The same device can display different information and change this information dynamically while in use.
When it comes to instrument panels, a basic distinction is made between the hybrid and freely programmable variants. Hybrid versions combine a number of indicator instruments with simple segment displays or small to medium display sizes. By contrast, the freely programmable instrument panels (FPIs) completely dispense with analogue indicators and a display fills the entire surface of the panel.
In the case of many end products – especially where vehicles are concerned – it is no longer just about functional properties. It is now increasingly a question of brand image too, combined with a defined value that is reflected in the design of all vehicle parts. The user interface is where users make direct contact with the product – and how this is experienced by users makes all the difference to the brand image. This interface is therefore more than simply a means of making functions available; it is increasingly becoming the communication channel between brand and consumer. GUIs make it easy to generate a brand image and create features that set it apart from competing products.
On a traditional instrument panel, for example, high value is represented by the colour of the background and instrument needles, bordering chrome rings and ambient background lighting. The first generation of FPIs that entered series production was still a digital reproduction of its analogue predecessor and important information such as speed, RPM and fuel level was presented by means of graphical indicators with almost identical quality. In future generations, however, new methods of presentation will find their way into vehicles with the GUI, with their design possibilities casting new light on the definition of quality.
Embedded SoC families from Fujitsu
In the automotive industry, it is important that the costs for hardware and software that are incurred when new technologies are implemented do not exceed those of existing solutions. Compatibility with available hardware and re-usability of software are important pre-requisites in this respect for keeping development costs low.
Fujitsu has defined its SoC (system-on-chip) families with the goal of being able to operate the different variants of hybrids and freely programmable instrument panels in vehicles (figure 1). Along with functionality and computing power, these families also take into account the difficult conditions of the automotive environment, such as electromagnetic compatibility, low power consumption, functional safety, data security and quality.
Figure 1: Segmentation for automotive instruments. For full resolution click here.
The entire product portfolio is based on a universal core architecture, which ensures compatibility among the individual controllers. With the Cortex cores from ARM, two families are defined that cover the full range of requirements and ensure complete inter-compatibility. The individual derivatives of these families are each SoC equipped with variable processing and graphical power and differing levels of integrated interfaces. The Cortex-R4 family, with 1.6 DMIPS/MHz, integrated memory, necessary system and network interfaces and SMC (stepper motor controllers), offers all the pre-requisites for hybrid instrument panels with up to six indicators and average display resolutions of 800 x 600 pixels, for example. With the integrated ‘IRIS’ graphics engine, 2D graphics operations are performed without placing any load on the CPU. Integrated interfaces such as I2C, I2S, CAN, USB and Ethernet provide the connection to other control devices and enable interaction with the user.
On the higher-end instrument panels, the display resolutions are higher, which means that more and more information in more complex graphics is shifted to the display. The computing power and graphical performance required for this are provided by the Cortex-A9 SoC from the ‘Emerald’ family, which integrates three display outputs, four video inputs and other interfaces such as I2C, I2S, CAN and USB, as well as the CPU and GPU. The GPU (graphics processing unit) in this instance comprises two processors that can carry out separate 2D and 3D operations. Here, the ‘IRIS’ 2D engine is combined with a 3D OpenGL core. Together with the layer technology used, individual image elements with different refresh rates can be computed according to the requirements of the application. Video streams in RGB888, ITU-656/601 format are read, superimposed and blended PiP (picture-in-picture) with computed graphics by means of the integrated video inputs. The ‘Emerald’ SoC units feature multiple display outputs.
The parameters of each of these outputs are user-programmable and support different resolutions, even those that deviate from defined standards such as 16:9 or 4:3. The parallel RGB signals can be output as TTL (transistor-transistor logic) levels or differentially via RSDS (reduced swing differential signalling). Each of the display outputs can manage up to eight layers, which can be either superimposed or blended on the display. A layer is a separate memory area the content of which can be displayed in differing resolutions, which means that the processes of computing individual image content can be de-coupled from each other. Since it is not necessary for the entire image to be re-computed at the maximum refresh rate when only partial elements change, this reduces the required computing power of the GPU.
Potential for developers
Such type of complex SoC brings benefits to developers by reducing the effort in various kinds. As many peripherals are integrated the required quantity of external devices is smaller. This makes the PCB layout much easier and the required PCB area is much smaller. Software developers avail on the standardisation of built in interfaces. These standards allow the reuse of available software parts and avoid, that the complete functionality needs to be re-developed. Software developer can further access to authoring tools, which are provided with such SoC and reduce the development time obviously.
Development costs reduced by complete systems
The switch to GUIs (graphical user interfaces) is something of an obstacle for many project managers and developers. Some of the know-how that has already been built up is rendered obsolete, and work must be done on complex graphics programming. This is a new challenge for semiconductor manufacturers. The expenditure associated with software development for an application rises as the complexity of the semiconductors increases and accounts for a significant portion of the overall costs. For the semiconductor manufacturers, this creates a need to offer additional software packages that extend far beyond the scope of driver layers and reduce the time and costs involved in developing the application.
In the form of FEAT (Fujitsu Embedded Solutions Austria GmbH), Fujitsu has excellent in-house expertise for developing and marketing comprehensive software solutions. Together with the development of the controllers, CGI Studio provided the first authoring tool for GUIs. The parallel development of hardware and software ensured that the interfaces were optimally co-ordinated with one another.
As a PC-based development environment, CGI Studio is the mediator between the professional design programs and the system environment, which supports user-friendly GUI programming for automotive applications (figure 2). 3D models pre-developed externally can be dragged and dropped into the application. The dynamics and animation can then easily be programmed and tested on the PC with the integrated scene composer. On the target hardware, the Canderra engine conveys the graphical commands required for the GUI generated to underlying software layers. The optimal connection to the GPU means that its available functionality and computing power can be utilised effectively without loading the CPU. CGI Studio provides full support for the integrated GPUs of all Fujitsu SoC families, brings down development times and thus significantly reduces software development costs.
Figure 2: Designing automotive GUI with the CGI Studio
It is with complete solutions like this that Fujitsu has made the leap from semiconductor manufacturer to system provider. Fujitsu will continue on the path it has taken towards helping to shape the technological revolution in vehicle display instruments.
360° view from inside the car
Along with GUIs, camera-supported assistant functions in vehicles are playing an increasingly important role. Cameras mounted on the sides of the vehicle provide a view of blind spots and thus improve safety (figure 3).
Figure 3: Higher safety camera-based driver assistance
The simple display enables each camera image to be shown independently on a defined area. Using the 360° Wrap-Around Video Imaging Technology developed by Fujitsu, a complete all-round view can be generated through spherical processing, which involves the individual camera images being placed over 3D geometry. Based on the position of the observer, the entire surroundings of the vehicle are shown in an image with a sense of perspective, with the angle of vision freely selectable, depending on the situation. This assistant function makes it easier for drivers to maintain their orientation and to safely recognise and correctly evaluate obstacles around the vehicle and their proportions. The current status of this technology can be seen in a demonstrator developed in collaboration with Magneti Marelli S.p.A. This application is based on the MB86R12 ‘Emerald P’, which is ideal for this application thanks to its four camera inputs.
About the author: Andreas Grimm is Senior PME at the Graphics Competence Center, Fujitsu Semiconductor Europe.
More information on https://emea.fujitsu.com/semiconductor
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
