[ad_1]
Increasing automotive complexity requires new approaches to in-car connectivity, especially the physical-layer interfaces that link sensors and displays to their associated electronic control units.
In the automotive industry, features such as advanced driver-assistance systems (ADAS), connected in-vehicle infotainment (IVI) and emerging autonomous driving systems (ADS) are more important than ever, making vehicles safer and improving the driving experience. Yet they also are creating new requirements that are increasing complexity and making product development more expensive and time-consuming.
Automakers are facing pressure to include the latest capabilities while containing costs, minimizing power consumption and ensuring electronic systems are reliable, safe and secure for the life of the vehicle. Meeting these expectations requires new approaches to in-car connectivity, especially the physical-layer interfaces that link sensors and displays to their associated electronic control units (ECUs).
Let’s take a look at these trends, the demands they create and ways to meet those requirements.
More cameras, more displays, more data
In 2021, the research firm Canalys estimated approximately one-third of new vehicles sold in major markets such as the U.S., Europe, Japan and China had ADAS features. It also predicted that half of all cars on the road in 2030 would be ADAS-enabled.
click for full size image
Figure 1: Vehicles with ADAS or ADS are equipped with a wide array of cameras and sensors for 360-degree visibility, raising new requirements for in-car connectivity. (Source: MIPI Alliance)
Advanced ADAS platforms can use up to 12 or more cameras — along with radar, lidar and ultrasonic sensors — for safety capabilities such as 360-degree visibility, lane-keeping assist, traffic sign recognition and automatic emergency braking. Many vehicles also have interior cameras to monitor driver alertness. Autonomous vehicles are equipped with even more sensors, and the requirements for resolution and performance are growing across the board.
The new electronic features also require more, larger and higher-resolution displays. Digital cockpits with electronic gauge clusters, head-up displays and virtual mirrors using cameras outside the vehicle are already common. Central stack, passenger-side and rear-seat displays are undergoing continuous improvement to take advantage of growing cloud-based sources of entertainment, navigation and local information.
The research firm IHS Markit has predicted that by 2026, 34.1% of new passenger vehicles will have digital instrument clusters, and 41% of central stack displays will measure 9 inches or larger. IVI displays commonly exceed 12 inches in size, with resolutions as high as 3840×2160 pixels and increasing.
Technology advances add requirements
The proliferation of these components raises issues around network performance, complexity and safety.
Connecting more components to processors — over links that may span the whole vehicle — increases network complexity. Traditionally, each sensor or display unit has its own wiring to its associated ECU, adding to the weight and manufacturing cost of the wiring harness. Meanwhile, increasing resolution and frame rates for image capture and display challenge OEMs to provide more bandwidth on each link without adding wiring.
With the growth of safety-critical use cases, automotive interfaces require functional safety features to meet industry requirements, such as compliance with the ISO 26262 standard to achieve Automotive Safety Integrity Levels (ASIL) B through D. Connections between onboard sensors and ECUs need to be protected under all conditions, throughout the vehicle’s life, to prevent faulty or missing data from causing driver or vehicle errors. The same is true of links from ECUs to displays used in applications such as video feeds from backup and parking assistance cameras.
The power of standardization
A standardized approach to sensor and display connectivity can help manufacturers meet these requirements. It can eliminate expensive, time-consuming integration and testing of proprietary solutions, which can delay the introduction of new features. Interoperability based on standard interfaces also allows more new players to enter the market, giving OEMs more suppliers to choose from.
MIPI Alliance offers such an approach with MIPI Automotive SerDes Solutions (MASS), an end-to-end framework for reliable, high-performance links with built-in functional safety (and security under development). MASS reduces complexity by allowing OEMs to implement common automotive interface protocols across a vehicle over A-PHY, the first standardized asymmetric long-reach SerDes physical-layer interface.
As the foundation of the MASS framework, A-PHY offers increasing performance and flexibility, along with the reliability and resiliency required for safety-critical applications. Its maximum downlink data rate has increased from 16 Gbps in A-PHY v1.0 to 32 Gbps per link in A-PHY v1.1, introduced earlier this year, with a roadmap to 64 Gbps and beyond. The maximum uplink data rate has also doubled in v1.1, from 100 Mbps to 200 Mbps. An ultra-low packet error rate of 10-19 and high noise immunity provide for reliable communication throughout a vehicle’s life.
A standardized interface such as A-PHY, with a reach of up to 15 m, can streamline integration and reduce network cost and complexity across a variety of architectures. It allows for links between edge components and ECUs anywhere in the vehicle, eliminating the need for “bridge” processors between short-reach interfaces (such as MIPI C-PHY or MIPI D-PHY) and a proprietary long-reach interface. A-PHY also allows manufacturers to daisy-chain multiple devices to an ECU over one cable.
Greater flexibility in A-PHY v1.1
A-PHY supports multiple existing interface protocols via adaptation layers and can be implemented in a variety of topologies and configurations. A-PHY v1.1 increases this flexibility to support more types of implementations.
click for full size image
Figure 2: The flexibility of A-PHY has increased in A-PHY v1.1 with support for Star Quad (STQ) cables. With three new configuration options, OEMs have more ways to meet connectivity requirements. (Source MIPI Alliance)
The addition of support for Star Quad (STQ) cables, a shielded cable with two differential pairs, in A-PHY v1.1 allows for three new configurations:
(Source MIPI Alliance)
- A dual-downlink configuration uses both pairs of conductors on an STQ cable to create the 32 Gbps downlink speed.
- An asymmetric option, with a 16 Gbps downlink and a 4 Gbps reverse downlink, enables high-speed data transfers in both directions, so one cable can serve pairs of devices such as a co-located camera and a display.
- A symmetric configuration provides a 16 Gbps downlink and a 16 Gbps reverse downlink over one STQ cable.
A-PHY v1.1 also allows manufacturers to migrate to A-PHY using legacy cables on existing platforms or lower-cost cable types on new platforms by making PAM4 encoding — which enables low-bandwidth, sub-1 GHz operation — available for the lower speed gears.
A-PHY has several built-in functional safety features, including cyclic redundancy checks (CRCs), an 8-bit message counter to detect packet loss and a timeout monitor to detect loss of communication. In addition, its unique retransmission scheme (RTS) recovers damaged packets for a steady connection, increasing noise immunity and minimizing packet errors. The MASS framework also includes additional features for functional safety in the upper layers of the protocol stack.
A time of opportunity
Standardized interfaces such as MIPI A-PHY, which will continue to advance with higher bandwidth and greater flexibility, enable manufacturers to embrace these possibilities while addressing new requirements for the connected, autonomous, shared and electric vehicle models of the future.
Raj Kumar Nagpal is Senior Staff Engineer at Synopsys, co-chair of the MIPI A-PHY Working Group, chair of the D-PHY Working Group and chair of the MIPI PHY Steering Group. |
Edo Cohen is Vice President, Ecosystem Development at Valens Semiconductor and co-chair of the MIPI A-PHY Working Group. |
Related Contents:
For more Embedded, subscribe to Embedded’s weekly email newsletter.
[ad_2]
Source link