Skip to main content

An affordable Cyber Biological System combining swarming biosensors and robotics

Periodic Reporting for period 1 - CyBioSys (An affordable Cyber Biological System combining swarming biosensors and robotics)

Reporting period: 2016-09-01 to 2018-08-31

The fundamental goal of this project was to demonstrate that mixed societies can benefit from advantages of both artificial and natural systems and outperforms their constituents. In this context, our objective was to design a cyber biological system (CBS) that interfaced a model organism of biological societies, the ants, with an artificial system based on robots developed for educational and research purposes, the Thymio robot. The main goal of this CBS was to explore the environment searching for resources and to exploit them as soon as discovered. To do so, the CBS benefited from the ability of the robots to quickly convey items and from the efficient exploration pattern of the ants to discover and evaluate potential resources. The synergy created by this collaboration outperforms pure artificial systems by saving the energy that should be consumed by the robots constantly exploring the environment and by delegating the detection of potential resource to animals. Similarly, the CBS surpassed pure biological systems by exploiting the resources more quickly, with the robots moving fast and straight to the target, and making the ants more rapidly available for further exploration. These improvements make the CBS a low energy-cost system that exploit more efficiently environmental resources than its components.

During the project, we identified the variation of the flow of workers at the entrance of the nest as a clear signal associated by the discovery of a food source by the ants. Then, we developed a device to monitor the behaviour of the ants and autonomously detect this signal. In parallel, we developed a mobile robot able to navigate in the presence of the ants and able to pick up a food source in the environment. Finally, we designed a binary choice experiment in which the ants had to discover and select the best food source available in the environment and the robot had to bring back the food source that has been selected by the ants. Thus, these results provide evidence for the possibility to transfer information from natural agents to artificial ones efficiently and the ability of artificial systems to facilitate the collective exploitation of environmental resources by animals.
WP1 Biological experiments

Identification of a signal associated with food discovery by ants: the flow of ingoing and outgoing ants at the nest entrance has been identified as a reliable indicator of the absence (low activity) or presence (high activity) of a food source in the environment.

WP2 Conception of the artificial components

Development of a mobile robot and a device to monitor the ants: we developed a sensor that tracks the motion of the ants in front of the nest entrance to detect an increase of activity. The number of moving pixels cumulated over 5-minute periods has been shown to be a reliable proxy to estimate the flow of ants. In parallel, we developed a mobile robot able to navigate in the presence of the ants and to follow them until it reaches the food source selected by the colony.

WP3 Experimentation with Cyber Biological Systems

Successful experiments with a mixed society of ants and artificial agents: we successfully combined the abilities of the natural and artificial agents to formed an efficient mixed society of ants and robots. This mixed group had to exploit the best resource available in the environment. To do so, the artificial system benefited from the ability of the ants to discovered and select the best resource in the environment while the ants benefited from the ability of the robot to convey the resource closer to the nest.

WP4 (M1 – M24) Management, dissemination & training

Dissemination of the results and Training of the ER: the results have been presented at five international scientific conferences and two science fairs. The ER has been interviewed by a national swiss radio as well as two journalists working for web magazine. A manuscript is currently in preparation to be submitted to a high impact factor journal. Along the project, the ER has been trained to numerous methods and techniques used in engineering (e. g. CAD software, computer vision or robotics).
We demonstrated that distributed natural and robotic systems each with specific task solving abilities can be integrated together to produce a synergetic system. These systems can be considered as a fundamental new class of distributed cyber biological systems able to solve tasks more efficiently than pure systems. In this project, the goal of the CBS was to discover environmental resources. However, our methodology could be applied to other tasks involving the exploration of a complex and unpredictable environment such as prospection or search and rescue (SAR). Currently, dogs used in SAR are deployed in a restricted zone and are worked by humans on foot. By coupling dogs with additional sensors such as GPS, gas detector or other sensors, the dogs could autonomously scan a larger area in a shorter time and in a safer way. In addition, the development of species-specific sensors could enlarge the number of animal species used in urban SAR to small animals like ferrets that can access restricted environment after building collapses for example. In parallel, the development of artificial systems coupled to natural ones is expected to create new solutions for current agriculture. Indeed, rather than developing automated solutions based on routines for animal care, it will allow the development of CBS able to adapt to the state of the animals. Sensors could detect sick animals more quickly by constantly checking their behaviour, still taking care of their nutrition as it’s already done for cows, for instance. Thus, such CBS could increase animal welfare without additional cost. On the other hand, such system will also provide new solutions against animals that have always been considered a threat to human welfare and agriculture. Indeed, rather than using preventive pesticides with a broad spectrum against animal pests, the monitoring of animal activity will allow targeted interventions with specific treatments.