[Paper Review] Implementation of Google Assistant & Amazon Alexa on Raspberry Pi.

Shailesh Arya, Samir Patel|arXiv (Cornell University)|Jun 15, 2020

IoT-based Smart Home Systems1 references4 citations

TL;DR

This paper presents a low-cost, dual-virtual-assistant smart speaker prototype built on a single Raspberry Pi 3, enabling concurrent use of Google Assistant and Amazon Alexa via voice-activated hotwords. By leveraging cloud-based NLP services for both assistants, the system processes voice input, interprets intent, and delivers spoken responses through a USB speaker, demonstrating seamless integration of two major voice ecosystems on a single embedded platform for home automation and general-purpose interaction.

ABSTRACT

This paper investigates the implementation of voice-enabled Google Assistant and Amazon Alexa on Raspberry Pi. Virtual Assistants are being a new trend in how we interact or do computations with physical devices. A voice-enabled system essentially means a system that processes voice as an input, decodes, or understands the meaning of that input and generates an appropriate voice output. In this paper, we are developing a smart speaker prototype that has the functionalities of both in the same Raspberry Pi. Users can invoke a virtual assistant by saying the hot words and can leverage the best services of both eco-systems. This paper also explains the complex architecture of Google Assistant and Amazon Alexa and the working of both assistants as well. Later, this system can be used to control the smart home IoT devices.

Motivation & Objective

• To develop a low-cost, single-hardware smart speaker prototype that supports both Google Assistant and Amazon Alexa.
• To enable users to invoke either assistant via distinct hotwords (e.g., 'Ok Google', 'Alexa') on the same device.
• To integrate both assistants with a shared Raspberry Pi 3 platform for voice-based control of IoT devices.
• To demonstrate the feasibility of running two cloud-dependent virtual assistants simultaneously on resource-constrained embedded hardware.
• To provide a DIY, open-source solution for home automation using widely available components.

Proposed method

• Deploying the Google Assistant SDK on Raspberry Pi 3 using the Google Assistant Library and OAuth2 authentication via a client secrets JSON file.
• Configuring the Alexa Voice Service (AVS) SDK on Raspberry Pi using the AVS Device SDK and setting up a security profile with a device-specific configuration.
• Using a USB microphone for voice input and a 3.5mm AUX-connected speaker for audio output, with both assistants processing voice commands via remote cloud-based NLP and intent recognition.
• Running both assistant services in parallel on the same Raspberry Pi, with separate processes listening for distinct hotwords ('Ok Google' and 'Alexa').
• Authenticating both assistants using device-specific OAuth2 credentials and refresh tokens obtained through web-based authorization flows.
• Integrating the system with IoT devices (e.g., smart plugs, bulbs) via native support in both Google Assistant and Amazon Alexa ecosystems.

Experimental results

Research questions

• RQ1 Can both Google Assistant and Amazon Alexa be concurrently operational on a single Raspberry Pi 3 without performance degradation?
• RQ2 How can two distinct cloud-based virtual assistants be integrated on a single embedded platform with shared audio input and output?
• RQ3 What are the key configuration and authentication steps required to enable both assistants on a Raspberry Pi for a DIY smart speaker?
• RQ4 To what extent can a low-cost, single-board computer like Raspberry Pi 3 handle real-time voice processing and dual-assistant interaction?
• RQ5 How can such a system be extended for practical home automation using standard IoT devices?

Key findings

• The system successfully runs both Google Assistant and Amazon Alexa in parallel on a single Raspberry Pi 3, with each assistant responding independently to its respective hotword.
• The integration was achieved using official SDKs from Google and Amazon, with no need for custom speech recognition models, relying instead on cloud-based NLP and intent processing.
• The setup process, including OAuth2 authentication and device registration, was successfully completed using standard developer tools and configuration files.
• The prototype demonstrated reliable voice interaction, with both assistants correctly interpreting commands and generating spoken responses through the USB speaker.
• The system enables control of standard IoT devices (e.g., smart plugs, bulbs) via voice commands through either assistant’s native ecosystem integrations.
• The entire solution can be built for significantly less cost than commercial smart speakers, making it a viable DIY alternative for home automation and voice interaction.