Method and system for self-driving labs

Abstract

Method for enabling and optimizing the use of a self-driving laboratory comprising a No-Code Cloud (NCC) platform and an automated robotic equipment; wherein said NCC platform communicates at least with a module of AI / ML algorithms to control said automated robotic equipment, to drive, orchestrate and execute operations, such as experiments, in full autonomy. The invention also relates to a system for self-driving laboratories comprising a No-Code Cloud (NCC) platform, a module of AI / ML algorithms and at least one robotic equipment; said NCC platform and experiment planners module being configured to control said robotic equipment.

Description

[0001] Method and system for self-driving labs

[0002] Field of invention

[0003] The present invention generally relates to autonomous or self-driving laboratories combining one or several automated robotic equipment(s) with artificial intelligence and machine learning (AI / ML) algorithms.

[0004] Prior art

[0005] Self-driving laboratories using Al algorithms are disclosed in the following references:

[0006] - ChemOS: an orchestration software to democratize autonomous discovery (ChemRxiv, March 7, 2018)

[0007] - ChemOS: Orchestrating autonomous experimentation (Science Robotics, June 20, 2018)

[0008] Description of the invention

[0009] The present invention provides an optimized method and system for using and enabling a selfdriving laboratory that includes at least one robotic equipment.

[0010] The method according to the invention comprises the remote connection of AI / ML -that serve as experiment planners- to the automated robotic equipment, via a network, and can optimize the experimental conditions or parameters in full autonomy.

[0011] The invention can be used in a wide variety of applications, such as experiments in research and development (R&D), process development and process scale-up, formulation, synthesis, catalysis across industries, including in chemistry, materials science, pharma, biotechnology, agrotechnology, fragrances, food, and energy. The method according to the invention comprises the following steps: a) Users connect remotely to a No-Code Cloud (NCC) platform from their computers, smartphones or tablets. The connection to the NCC may be established using a Graphical User Interface (GUI), an Application Programming Interface (API) or a Software Development Kit (SDK). b) Users define their experiments, including input parameters and measurements or objectives. c) Users select the AI / ML algorithm (experiment planner) that will suggest experimental conditions for the first iteration of experiments. d) Users define the means of communication to the automated or robotic equipment present in their laboratories. e) Users start experiments. f) The experiment planner suggests experimental conditions, which are sent automatically to the device(s) controlling automated or robotic equipment through the network. g) The results of the experiment, or measurements, are automatically collected by computing device, stored in digital databases in the NCC platform and used to retrain AI / ML algorithms so that these suggest the new values for each parameter, that is, the best set of experimental conditions to use in the next iteration of experiments, and submits these instructions automatically to the control device.

[0012] The invention provides several advantages and original features with respect to the prior art, in particular:

[0013] - A NCC platform that communicates with a module of AI / ML algorithms, named “experiment planners”, and preferably with a database(s) module to control automated robotic equipment(s), to drive, orchestrate and execute operations, such as experiments, in full autonomy.

[0014] The NCC platform acts as a workflow manager and as a queuing system.

[0015] The experiment planners module may advantageously include trial-and-error methods, design of experiment methods, grid search methods, AI / ML algorithms. The database(s) module comprises data from users that are automatically updated as each experiment is being optimized by the experiment planners module.

[0016] The network allows the transfer of experimental conditions from computing device(s) to control the robotic equipment(s), including for instance RESTful application programming interfaces (API), software development kits (SDKs) and file transfer protocols.

[0017] The network automatically translates the set of conditions from experiment planners to the format required by the control of the robotic equipment(s).

[0018] One or several equipment(s) may be simultaneously controlled.

[0019] The system according to the invention may be configured in different manners.

[0020] Some configurations are briefly presented hereafter.

[0021] Configuration 1 (figure 1)

[0022] A NCC platform communicates with the “experiment planners” module to control at least one automated robotic equipment, to drive, orchestrate and execute operations, such as experiments, in full autonomy.

[0023] Configuration 2 (figure 2)

[0024] This configuration is identical to configuration 1 but with the addition of a module including at least one database to digitize users' experimental result(s).

[0025] Configuration 3 (figure 3)

[0026] This configuration is identical to configuration 2 but with the addition of a data processing module to interpret experimental results.

Claims

Claims1. Method for enabling and optimizing the use of a self-driving laboratory comprising a No-Code Cloud (NCC) platform and an automated robotic equipment; wherein said NCC platform communicates at least with a module of AI / ML algorithms to control said automated robotic equipment, to drive, orchestrate and execute operations, such as experiments, in full autonomy.

2. Method according to claim 1 wherein said NCC platform also communicates with a database(s) module to digitize users’ experimental results.

3. Method according to claim 1 or 2 wherein said NCC platform also communicates with a data processing module to interpret experimental results.

4. System for self-driving laboratories comprising a No-Code Cloud (NCC) platform, a module of AI / ML algorithms and at least one robotic equipment; said NCC platform and experiment planners module being configured to control said robotic equipment.

5. System according to claim 4 furthermore comprising a database(s) module.

6. System according to claim 4 or 5 furthermore comprising a data processing module.

Images