Skip to content

DA2. Agents and Environments

Statement

For each of the following agents, describe all the possible PEAS (Performance, Environment, Actuators, and Sensors) and then identify the environment type.

  • Self-vacuum cleaner
  • Speech recognition device
  • Drone searching for a missing person
  • Computer playing chess
  • Alexa/Siri performing as a smart personal assistant.

Answer

Introduction

An agent is anything that perceives its environment through sensors and acts upon it through actuators (Poole & Mackworth, 2017). The text will start by defining the PEAS components and then describe the PEAS for each agent in the prompt and conclude with a brief summary about those agents.

PEAS is a famous framework in AI that divides the agent’s design into 4 components. Performance defines the measures of success, or describes the state of world when the purpose (goal) of the agent is achieved. Environment is the external context in which the agent operates. Actuators are the mechanisms through which the agent can act upon the environment. Sensors are the mechanisms through which the agent can perceive the environment (GeeksforGeeks, 2020).

The types of environments includes fully vs partially observable, single-agent vs multi-agent, deterministic vs stochastic, discrete vs continuous, known vs unknown, and static vs dynamic (Aouissi, n.d.).

Self-vacuum cleaner

The Performance of a self-vacuum cleaner can be measured by the cleanliness of the room (keep the room clean), efficiency (use less energy and move faster), safety (avoid obstacles and accidents), and mobility (move around or climb stairs). The Environment includes the house and its interiors. The Actuators include wheels, vacuum, and brushes. The Sensors include cameras, dirt detectors, cliff sensors, bump sensors, and wall sensors.

The environment is dynamic (environment changes as other agents can move around such as pets or humans), partially observable (agent can only see up to the limits of its sensors), multi-agent (other agents can be present), stochastic (the environment is unpredictable as other actors can add more dirt or obstacles), continuous (the agent can move in any direction), and unknown (the designer of the agent does not know the exact layout of the house before hand, and layout can change).

Speech recognition device

The Performance of a speech recognition device can be measured by the accuracy of the speech recognition, speed of recognition, and the ability to understand different accents. The Environment includes the device physical and the person speaking to the device. The Actuators include the speakers, or text output. The Sensors include microphones, cameras, and text input readers.

The environment is static (does not change), fully observable (the agent can observe the entire environment), single-agent (agent does not interact with other agents except the person), deterministic (it only depends on the words of the person), discrete (the agent can only recognize a finite number of words), and known (the designer of the agent knows the targeted language before hand).

Drone searching for a missing person

The Performance of a drone searching for a missing person can be measured by the efficiency (finding the person quickly), accuracy (covered areas are thoroughly inspected). The Environment includes the area where the person is missing (a place or a city). The Actuators include the engines, light projectors, communication devices to the base, and other hardware. The Sensors include cameras, recognition devices, GPS, microphones, and other sensors.

The environment is dynamic (the person can move around), partially observable (the drone can only see up to the limits of its sensors), multi-agent (other agents can be present), stochastic (the environment is unpredictable as the person can move around), continuous (the agent can move in any direction), and unknown (the designer of the agent does not know the exact location of the person before hand).

Computer playing chess

The Performance of a computer playing chess can be measured by the ability to win the game, and the speed of making decisions. The Environment includes the chessboard and the pieces. The Actuators include the chess pieces. The Sensors include cameras, chessboard readers, and other sensors.

The environment is static (does not change), fully observable (the agent can observe the entire environment), single-agent (agent does not interact with other agents), deterministic (it only depends on the moves of the opponent), discrete (the agent can only make a finite number of moves), and known (the designer of the agent knows the rules of chess before hand).

Alexa/Siri performing as a smart personal assistant

The Performance of Alexa/Siri can be measured by the ability to understand the user’s commands, the ability to provide the correct information, and the speed of providing the information. The Environment includes the device and the user. The Actuators include the speakers, text output, and programs that are necessary to communicate with the OS to open applications or do calls. The Sensors include microphones, cameras, and text input readers.

The environment is static (does not change), fully observable (the agent can observe the entire environment), single-agent (agent does not interact with other agents), deterministic (it only depends on the words of the user), discrete (the agent can only recognize a finite number of words), and known (the designer of the agent knows the targeted language before hand).

Conclusion

To conclude, the PEAS framework is a useful tool to understand the design of agents, and here are two tables that summarize the PEAS for each agent and the environment type.

Agent Performance Environment Actuators Sensors
Self-vacuum cleaner Cleanliness, efficiency, safety, mobility House, interiors Wheels, vacuum, brushes Cameras, dirt detectors, cliff sensors, bump sensors, wall sensors
Speech recognition device Accuracy, speed, understanding accents Device, person Speakers, text output Microphones, cameras, text input readers
Drone searching for a missing person Efficiency, accuracy Area where the person is missing Engines, light projectors, communication devices, hardware Cameras, recognition devices, GPS, microphones, sensors
Computer playing chess Ability to win, speed Chessboard, pieces Chess pieces Cameras, chessboard readers, sensors
Alexa/Siri Ability to understand, provide correct information, speed Device, user Speakers, text output, programs Microphones, cameras, text input readers
Agent Dynamic vs Static Fully vs Partially Observable Single vs Multi-Agent Deterministic vs Stochastic Discrete vs Continuous Known vs Unknown
Self-vacuum cleaner Dynamic Partially Observable Multi-Agent Stochastic Continuous Unknown
Speech recognition device Static Fully Observable Single-Agent Deterministic Discrete Known
Drone searching for a missing person Dynamic Partially Observable Multi-Agent Stochastic Continuous Unknown
Computer playing chess Static Fully Observable Single-Agent Deterministic Discrete Known
Alexa/Siri Static Fully Observable Single-Agent Deterministic Discrete Known

References

‌ ‌