A prototype of having a meaningful conversation with non-talkative objects.
You love your garden but your garden doesn’t love you back. You try to do the right thing, but what is the right thing? A garden needs care and understanding. But not everyone is an expert, not everyone knows what a garden needs. So, why not ask?
What if your garden could talk to you?
A garden is perceived a thing, but what if we could turn it into an entity? Giving the garden a voice and conversational skills, connected to knowledge and given sensoring abilities.
Then you could talk to her through the social messaging apps you already use on a daily bases like Facebook messenger, Skype or Telegram. Or talk with your garden using Google Home while doing the dishes.
How can we change this rather simple concept of taking care of your plants and take it to the next level?
How we did it
Based on Artificial Intelligence a chatbot is a powerful tool to support building smart conversations and trusted relationships. For example:
- Your garden indicates what it needs. Advice is given based on a database of information about plants, and real-time information and feedback of your plants using sensors. And if your plants are located outside, weather can also be a factor to take into account.
- Your garden can ask you to trim a specific shrub in your garden based on time of the season. Instructions can also be provided if the user asks for support.
By using various technologies, such as sensors, cloud computing and chatbot services, we were able to make a working Proof-of-Concept.
The way we achieved this is by using plant sensors from Xiaomi called the Flower Care Smart Monitor. This device uses various sensors to retrieve information about the plant, giving us the moisture and fertility rates of the plant, the temperature of its environment and how much sunlight it gets. This data will then be pulled from the sensor using Bluetooth Technology, and with the help of a Raspberry Pi computer the data will be sent into the cloud for further processing.
The cloud partner that we have chosen for this job is Amazon Web Services (AWS). AWS provides all the infrastructure needed for this to work. The AWS components used are; AWS IoT, AWS DynamoDB, AWS Lambda and AWS API Gateway. However, Microsoft Azure and Google Cloud Platforms both provide similar services to make this concept work.
High-level architecture of our Proof-of-Concept.
For the chatbot service, we use Dialogflow. Dialogflow is a Natural Language Processing (NLP) chatbot service provided by Google. This chatbot service can easily be linked to chat clients such as Telegram, FaceBook Messenger or it can be connected to a Google Assistant so you can talk to the plants using your voice.
To start the process we send the data that has been pulled from the sensors to the AWS IoT service and store it into a AWS DynamoDB database. Every time new data is added an AWS Lambda Function will be triggered. This function checks the well-being of the plants and notifies the user by sending messages if anything is wrong, for example: not enough water or sunlight available, too cold, etc.
And after receiving a status message, the user is able to continue the conversation by chat. While the message is being sent to the user, the context is also being sent to the chatbot service. Without this step, the chatbot will be unable to know what the conversation is about. Even when the user directly replies with a message corresponding to the subject. When the chatbot knows the context of the conversation, it can reply with an appropriate answer.
To clarify: the user can ask the chatbot questions like, "How is the Aloe Vera doing?" or "How do I take care of the Aloe Vera?". The chatbot will then know what the context is of the conversation and stick to it.
With all of this in mind, we created the possibility of having a conversation with a garden that is mutually beneficial. Becoming a trust worthy help for consumers and even offer products based on their needs.
But this doesn't only apply to plants. This concept of having a meaningful conversation with non-talkative objects can be applied to everyday things such as groceries, or even clothes. Nothing is off-limits.
Want to stay posted about AI for your industry?
Want to know more, or stay in touch with our innovations? Follow our blog and if you have specific questions about this used case, please get in touch with Dimitri Teunissen, Senior Creative Consultant, firstname.lastname@example.org
Looking for an inspiring Computer Science Internship? Have a look here.