LAVA: The exciting convergence of Android SOA and ADAS for SDV

LAVA unlocks the full potential of Software Defined Vehicle by bridging ADAS and Android SOA concepts

LAVA: The exciting convergence of Android SOA and ADAS for SDV

LAVA unlocks the full potential of Software Defined Vehicle by bridging ADAS and Android SOA concepts

LAVA: The exciting convergence of Android SOA and ADAS for SDV

LAVA unlocks the full potential of Software Defined Vehicle by bridging ADAS and Android SOA concepts

LAVA: The exciting convergence of Android SOA and ADAS for SDV

LAVA unlocks the full potential of Software Defined Vehicle by bridging ADAS and Android SOA concepts

It comes as no surprise that the seamless integration of ADAS and Android infotainment plays a pivotal role in establishing a cohesive and intelligent ecosystem for software-defined vehicles (SDVs). Currently, the lack of such integration poses a significant hurdle for the automotive industry, preventing it from reaping the benefits of synergistic development and capabilities.

The potential benefits are vast and varied. They encompass sophisticating of car-centric IVI (In-Vehicle Infotainment) applications that could leverage radars and sensors for augmented reality (AR) overlays, employ the driver monitoring system camera for video calls, incorporate autopilot computer vision into the main Head Unit, or harness ADAS AI cores for IVI applications. Furthermore, this integration holds the key to unleashing the potential for third-party ADAS software development driven by the IVI demands.

SOA in Android Automotive, motivation

The widespread adoption of service-oriented architecture (SOA) in automotive software development has undeniably facilitated the design, development, and maintenance of complex systems. Simultaneously, automakers are rapidly embracing Android due to its extensive app ecosystem, strong market demand, and the potential to address IVI fragmentation. However, Android comes with a proprietary IPC mechanism – binder, introducing yet another communication protocol which additionally adds to heterogeneity of the vehicular system. Automotive stakeholders seek unambiguous automotive friendly AOSP alternatives to accelerate development cycles for new applications. They often choose collaboration with Google through Google Automotive Services (GAS) and demand a strong differentiation of Android Automotive from other Android flavors to unlock true car-centric apps and pave the way to a centralized architecture. Nevertheless, the market currently lacks a comprehensive and impactful solution for intelligent inter-operability between the available automotive SOA and Android over binder while effectively managing safety-critical operations.

Amidst the market void, LAVA (Linking ADAS Via Android) emerges as a comprehensive solution that combines techniques, tools, and automation to generate Android APIs for seamless access to ADAS functions. LAVA offers a scalable and platform-agnostic implementation of the ADAS-Android IVI integration, promoting the harmonious coexistence of Android binder and conventional automotive SOA concepts while effectively managing safety-critical operations.

LAVA as the solution

The goal behind LAVA is to establish this technological bridge by imposing an IDL (Interface Definition Language) alignment between ADAS OS and Android and overcoming Android (Automotive) SOA constraints while forcing generalization – using code generation and reducing custom integration to a bare minimum. For the exchange of data coming from hardware resources and software components between ADAS and Infotainment LAVA embraces SOME/IP and DDS as SOA mechanisms for signal messages and smaller amounts of data. For big data streams, the solution provides DataChannelAPI which can be used in any component or environment and covers streaming via TCP, UDP, RTP, RTSP and AVB protocols. For the translation between the IDLs (ARXML – AUTOSAR XML, AIDL – Android interface definition language, Franca IDL, IDL), MDD (Model Driven Development) has been adopted as a development of choice due to its facilitated implementation, validation, verification, and maintenance through EMF (Eclipse Modeling Framework).

Challenges

There are several significant challenges involved in LAVA. The first challenge involves adopting the SOME/IP and DDS communication mechanism paradigm while transitioning from AIDL’s “binder-ized” communication for the effective distribution of SOME/IP data to Android applications. Another challenge pertains to enabling the exposure of ADAS information in Android services deployed in both service and vendor partitions, i.e. simultaneous availability in the application layer and HAL modules. Additionally, there is the challenge of implementing meta models for AIDL and IDL to comply with the principles of MDD. Finally, ensuring safety and security aspects of the solution pose additional hurdles to overcome.

The key to safe implementation of the bridge lies in one-on-one compatibility with Adaptive AUTOSAR, the techniques, and tools used in the solution. As SOME/IP communication mostly focuses on efficient data exchange rather than safety mechanisms, and while it can be used safely inside ADAS, it is crucial to implement the necessary safety measures and adhere to relevant safety standards and guidelines to ensure the overall safety of the integration.

Where is LAVA currently – communication stack, tool, automation

LAVA solution currently supports widely adopted SOME/IP and DDS protocols and offers easy scalability and expansion options with additional IPC mechanisms. The communication stack seamlessly adapts binder, SOME/IP and DDS protocols diversity, and implements various communication types such as event-triggered, on-demand, RPC (Remote Procedure Call), and others. The stack incorporates security enhancements primarily focused on SOME/IP. Additionally, it supports Android services for both vendor and system partitions, while providing compatibility with Native, Java, and Kotlin programming languages for their development.

A particularly fascinating aspect of LAVA lies in its automation capabilities. By utilizing EMF, meta models for AIDL and IDL definition languages are created, enabling implementation of seamless model-to-model translation between the automotive industry’s most commonly used interface definition languages, including ARXML, AIDL, FIDL, and IDL. LAVA meticulously manages semantic mapping and navigates the differences in paradigms utilized by each language. In addition to model translation, the LAVA tool facilitates the generation of a crucial part of the communication stack. Based on the ADAS systems modeled via ARXML, the tool enables the creation of an Android service with integrated SOME/IP or DDS clients in the IVI domain. This facilitates seamless data transmission with ADAS and allows for the distribution of information throughout the entire Android system via binder. What is even more remarkable is that all of this can be achieved with the LAVA tool in just a few clicks. A simple GUI is available for Windows and Linux.

LAVA at the forefront of the path towards embracing SDV

LAVA goes in line and one step ahead of the adoption of the SDV in the Android ecosystem. While the Android community recognizes the need for addressing platform heterogeneity and consolidation, and connectivity as the solution to faster automotive development cycles and optimal integration, as of today little has been done in Android SOA for a full-fledged SDV development and better utilization of resources through cross-silo bridge between ADAS and IVI. In the realm of LAVA, the general aspiration to support co-existence with currently adopted protocols in automotive manifests as provisioning of inter-operability between binder and the protocols adopted in automotive (SOME/IP and DDS). Where general intentions of the community stress automation as a fundamental code generation presumption, LAVA offers tools for generating solutions and translations between common IDLs. Subsequently, as general plans presume support for often-used safety OS solutions, we at LAVA emphasize one-on-one compatibility with Adaptive AUTOSAR, the future-proof basis for automotive ECUs (Electronic Control Unit) and the standard for computing-intensive task in the automotive.

For more information, visit LAVA page, or watch LAVA in motion including demos on Qualcomm and Telechips.

You may also like