Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
The Alexa Auto SDK is for automotive OEMs to integrate Alexa directly into vehicles. **v4.0.0** * v4.0.0 released on 2021-12-15 **Enhancements** * Deprecated the C++ and Java platform interfaces in favor of an asynchronous message-based API. Auto SDK client applications use the new `MessageBroker` to publish and subscribe to Alexa Auto Services Bridge (AASB) messages. The C++ sample app is refactored to use the new API to provide a reference implementation for Linux platforms. The Alexa Auto Client Service (AACS) sample app provides the reference implementation for Android platforms. See the [Auto SDK Migration Guide](./MIGRATION.md) for help migrating your application to use the new API. * Enhanced the Auto SDK build system with the Conan package manager. The new build system introduces modular builds, better dependency management, and simpler build artifacts. The Auto SDK build system includes the Auto SDK Builder Tool script, which wraps the Conan build commands with a simple interface similar to the previous version of Auto SDK Builder. See the [Build Alexa Auto SDK documentation](./BUILDING.md) for details about the build system and the [Migration Guide](./MIGRATION.md) for help migrating your build to the new version of Builder Tool. * Extended the features of Alexa Presentation Language (APL) support for automotive. The `APL` module provides messages to report vehicle properties such as the display theme, driving state, and ambient light conditions. The property settings affect how APL documents render on screen; for example, some APL content is automatically hidden when the vehicle starts moving, and the display contrast updates with the day or night mode setting. Auto SDK 4.0 supports APL 1.9. For more information about the Auto SDK `APL` interface, see the [APL module documentation.](./modules/apl/README.md) * Added the `CustomDomain` interface, which establishes a bidirectional communication channel between your Auto SDK client application and your custom cloud skill. `CustomDomain` includes messages for exchanging directives, events, and context between the vehicle and your skill, achieving a fully customizable experience. For more information about the Auto SDK `CustomDomain` interface, See the [Custom Domain module documentation.](./modules/custom-domain/README.md) * Added the `MediaPlaybackRequestor` interface, which enables Alexa to play the user’s favorite media content as soon as they start their vehicle. `MediaPlaybackRequestor` simplifies content selection for the user by removing the need for the user to use buttons or voice commands to resume the Alexa media content that was playing when they stopped the vehicle. For more information about the Auto SDK `MediaPlaybackRequestor` interface, See the [Alexa module documentation.](./modules/alexa/README.md) * Extended the `AudioOutput` interface and added configuration to allow ducking Alexa media. Your application can use this feature for enhanced control of Alexa content audio focus according to your platform requirements. For more information about audio ducking, see the [Core module documentation.](./modules/core/README.md) * Updated the Auto SDK to use AVS Device SDK Version 1.25.0. For information about this version of AVS Device SDK, see the [AVS Device SDK release notes.](https://developer.amazon.com/en-US/docs/alexa/avs-device-sdk/release-notes.html#version-1250) * Added LVC support for Alexa Custom Assistant specialized handoffs. You can configure the default fallback and self-introduction prompts for your custom assistant while offline. For more information, see the `Alexa Custom Assistant` extension documentation. * Integrated the Auto SDK Conan build system enhancements to AACS and the AACS sample app. You can use a single Gradle command to build AACS and the AACS sample app. For build instructions, see the [AACS documentation.](./aacs/android/README.md) * Added the following enhancements to the AACS sample app: * **Additional languages—** The AACS sample app supports the following languages: *US English* (`en-US`), *Australian English* (`en-AU`), *Canadian English* (`en-CA`), *Indian English* (`en-IN`), *British English* (`en-GB`), *German* (d`e-DE`), *Spanish* (`es-ES`), *Mexican Spanish* (`es-MX`), *US Spanish* (`es-US`), *French* (`fr-FR`), *Canadian French* (`fr-CA`), *Hindi* (`hi-IN`), *Italian* (`it-IT`), *Japanese* (`ja-JP`), and *Brazilian Portuguese* (`pr-BR`). The sample app language setting matches the device’s system language setting and syncs the with Alexa as long as the setting is in the supported language list. If Alexa does not support the system language, the sample app GUI defaults to en-US and presents a list of languages for the user to choose from. Once the user selects the language override, the system language does not sync with the sample app again until the user logs out or disables Alexa. * **Network error prompts—** You can configure the sample app to provide feedback to the user when Alexa cannot respond due internet connection issues. The feedback is a voice prompt or an error screen depending on the user action. * **Alexa app assets—** The sample app can show Alexa logos (assets) on the setup screen and display cards instead of showing placeholder assets. * **Comms UI improvements—** Updated the contacts uploading logic in the `Comms UI` AACS app component to ensure the sample app only uploads the contacts for the primary phone. * Updated the AACS Telephony library to get the outgoing phone account using the Android standard API `getDefaultOutgoingPhoneAccount`. AACS Telephony no longer sends an account query intent when receiving the `PhoneCallController.Dial` message from the Auto SDK Engine. * Added a new intent `com.amazon.aacstelephony.bluetooth.connectionCheckCompleted`, which AACS Telephony service broadcasts when it finishes the initial bluetooth connection check. * Updated the `alexa-auto-lwa-auth` app component to use the `Authorization` Auto SDK interface for CBL authorization. **Other changes** * Moved several source code directories within the `aac-sdk` root directory to support the enhanced build system. * Removed `aac-sdk/platforms/android/`. The deprecated Java platform interfaces and JNI are in their respective modules. For example, the Alexa module Java interfaces and JNI are moved from `aac-sdk/platforms/android/modules/alexa/` to `aac-sdk/modules/alexa/android/` * Removed `aac-sdk/extensions/aasb/` because using AASB messages with MessageBroker is the primary Auto SDK API. AASB code for each module is in the respective module directory. For example, the AASB code for the Alexa module is in `aac-sdk/modules/alexa/aasb/`. Note that the AASB message headers to include in your application are not in this directory since they are generated as part of the Auto SDK build output. * Moved `aac-sdk/extensions/system-audio/` to `aac-sdk/modules/system-audio/` * Moved `aac-sdk/extensions/bluetooth/` to `aac-sdk/modules/bluetooth/` * Moved `aac-sdk/extensions/loopback-detector/` to `aac-sdk/modules/loopback-detector/` * Moved `aac-sdk/platforms/android/alexa-auto-client-service/` to `aac-sdk/aacs/android/` * Moved `aac-sdk/platforms/android/alexa-auto-client-service/app-components/` to `aac-sdk/aacs/android/app-components/` * Moved `aac-sdk/samples/android-aacs-sample-app/` to `aac-sdk/aacs/android/sample-app/` * Moved `aac-sdk/platforms/android/alexa-auto-client-service` `/commonutils/` , `/ipc/`, and `/constants/` to `aac-sdk/aacs/android/common/` * Moved AACS media player files to a directory `audioOutput` within `aac-sdk/platforms/android/alexa-auto-client-service/service/` * Moved the Media App Command and Control Android library from `aac-sdk/platforms/android/maccandroid/` to `aac-sdk/aacs/android/service/modules/maccandroid/` * In the LVC extension, the `LocalSearchProvider` AASB messages now have topic `LocalNavigation`. For example, the existing message `LocalSearchProvider.SearchRequest` in 3.3 is `LocalNavigation.SearchRequest` in 4.0. The next major release version of Auto SDK will change the topic back to `LocalSearchProvider`. * Deprecated the option to build AACS as an APK. Starting from Auto SDK 4.0, you can only build AACS as an AAR. * Removed the Android sample app based on the Java platform interfaces. The AACS sample app demonstrates using Auto SDK on Android. **Resolved Issues** * Fixed an issue preventing the generic `DEFAULT` type `LocalMediaSource` from working in offline mode with LVC. * Fixed a race condition in `SpeechRecognizer` in which enabling wake word detection immediately after calling `startCapture()` resulted in a missing call to `stopAudioInput()` when wake word detection was later disabled. * Fixed a deadlock that could occur in an application that uses the deprecated `AuthProvider` interface and starts, stops, and restarts the Engine in quick succession. * Fixed an issue in which Spotify playback commands were delayed on QNX. * Fixed an issue in which the Engine added malformed `PhoneCallController` context to `PhoneCallController` events sent to Alexa. * Fixed an issue in which AACS did not acquire audio focus prior to playing Alexa speech. **Known Issues** **General** * If you do not specify the `deviceSettings.locales` field of the Alexa module configuration, the Engine automatically declares support for the following locale combinations: ["en-US", "es-US"], ["es-US", "en-US"], ["en-IN", "hi-IN"], ["hi-IN", "en-IN"], ["fr-CA", "en-CA"], ["en-CA", "fr-CA"]. The Engine does not automatically declare support for default locale combinations if you assign an empty value to the `locales` field. * The Engine does not persist the `aace.alexa.wakewordEnabled` Engine property setting across device reboots. Your application has to persist the setting and set the property again at each Engine start. AACS implements persisting this property and hence does not have this issue. * If your Linux platform does not use AVX2 instructions, the Amazonlite wake word library initialization causes an illegal instruction error. * When using LVC and stopping the Engine, the `AlexaClient` connection status remains `CONNECTED` because the connection to LVC is not disabled. Your application should not accept user utterances while the Engine is stopped despite the connection status showing `CONNECTED`. * The [Alexa Automotive UX guidelines](https://developer.amazon.com/en-US/docs/alexa/alexa-auto/display-cards.html#dismiss-display-cards) specify when to automatically dismiss a `TemplateRuntime` display card for each template type. The Engine publishes the `TemplateRuntime` interface messages `ClearTemplate` and `ClearPlayerInfo` based on the timeouts configured in the `aace.alexa.templateRuntimeCapabilityAgent` Engine configuration. However, the configuration does not provide enough granularity to specify timeouts for different types of display cards. Consequently, there is no way for your application to configure automatically dismissing local search templates (e.g., `LocalSearchListTemplate2`) with a different timeout than other templates (e.g., `WeatherTemplate`). The configuration also does not provide a way for you to specify infinite timeout for `NowPlaying` cards. You must implement your application’s dismissal logic for display cards and media info accordingly. * When the user requests to view their list of timers on an APL-enabled application, they cannot use an utterance such as “Alexa, scroll up” to scroll through the list shown on the APL card. * There is a rare race condition in which publishing the `AlexaClient.StopForegroundActivity` message does not cancel the active Alexa interaction. The race condition can happen when the application publishes the message at the beginning of the `THINKING` state `AlexaClient.DialogStateChanged` transition. * On the Poky Linux 32-bit platform, the C++ sample app shuts down with an error on launch. * In offline mode with LVC, you might not see the `AlexaClient.DialogStateChanged` `THINKING` state transition if the user invokes Alexa with hold-to-talk and your application provides the audio input data in one large chunk. * In offline mode with LVC, Alexa gets stuck in the `THINKING` state and does not respond after changing the locale setting. The state recovers after a few minutes. * The `CBL` module uses a backoff when refreshing the access token after expiry. If the internet is disconnected when the Engine attempts the refresh, it might take up to a minute to refresh the token after the internet connection is restored. * Some `Core` module messages published by the Engine do not have a corresponding message for the application to report a handling failure. For example, if the user invokes Alexa by tap-to-talk, and the application cannot handle the `AudioInput.StartAudioInput` message, the Engine assumes the application handled the message properly and will provide audio data. As a result, the Engine state and application state might become out of sync. The affected messages are the following: * `AudioInput`: * `StartAudioInput` * `AudioOutput`: * `SetPosition` * `VolumeChanged` * `MutedStateChanged` **Car control** * If you configure the Auto SDK Engine and connect to Alexa using a set of endpoint configurations, you cannot delete any endpoint in the set from Alexa. For example, after you configure set A with endpoints 1, 2, and 3, if you change your car control configuration during development to set B with endpoints 2, 3, and 4, Alexa retains endpoint 1 from set A, which might interfere with resolving the correct endpoint ID for your utterances. However, any endpoint configurations with matching IDs override previous configurations. For example, the configuration of endpoint 2 in set B replaces endpoint 2 in set A. During development, limit configuration changes to create only supersets of previous endpoint configurations. Work with your Solutions Architect or Partner Manager to produce the correct configuration on the first try. **Communications** * Alexa does not understand DTMF utterances that include letters. For example, "press A" and "dial 3*#B" do not result in the correct DTMF directives. * The user might experience unexpected results by trying to dial or place calls in the following ways: * Using utterances that include “double”, “triple”, “hundred”, or “thousand.” For example, calling a number such as 1-800-xxx-xxxx by saying “Alexa call one eight *double oh*...” * Pressing special characters such has “#” or “*” by saying "Alexa press * #." * The user cannot accept or reject incoming Alexa-to-Alexa calls by voice while playing a skill with extended multi-turn dialogs, such as Jeopardy or Skyrim. **Entertainment** * If the user requests Alexa to read notifications while music is playing, they might hear the music play for a split second between the end of one notification and the start of the next. * When an external media player authorization is in progress during Engine shutdown, a rare race condition might cause the Engine to crash. * If your application cancels an Alexa interaction by sending the `AlexaClient.StopForegroundActivity` message to the Engine during music playback, the Engine might erroneously request your application to dismiss the` NowPlaying` media info by publishing the `TemplateRuntime.ClearPlayerInfo` message. Your application should not dismiss the media info in this scenario. * When using the `System Audio` module, Audible and Amazon music might not play correctly on i.MX8 boards. **Local search and navigation** * In offline mode with LVC, after the user requests a list of POIs with an utterance such as “Alexa, find a nearby Starbucks”, Alexa does not recognize followup requests such as "Alexa, select the first one" and does not display or read detailed information about the requested selection. **AACS** * If you do not use the default audio output implementation (i.e., your application handles `AudioOutput` AASB messages), your application will not receive the `AudioOutput.Stop` message if Alexa media is playing when AACS shuts down. As a workaround, your application can listen to `AASB.StopService` or adopt `AACSPinger` to listen to the `STOPPED` state of AACS and stop the media accordingly. **AACS Sample App** * The AACS Sample App does not show the language selection screen when the app is built with Preview Mode. * The AACS Sample App only shows the language selection screen if there is a language mismatch with the system language setting at the first app launch. [Read the SDK Docs](https://alexa.github.io/alexa-auto-sdk/)
- Loading branch information
1 parent
2607c7b
commit 9324a77