When robots employ intelligent phenomena to solve a problem, such phenomena are collectively called “robot intelligence.” This volume presents the latest advancements in robot intelligence technology, which adapts cognitive architectures that model human cognition for use in robots. The book is a collection of 101 papers, the proceedings of the First Conference on Robot Intelligence Technology and Applications.

Robot intelligence is divided into six classes, which are covered in this book as follows: cognitive intelligence, social intelligence, and behavioral intelligence are covered in chapter 1, and ambient intelligence, collective intelligence, and genetic intelligence are covered in chapter 2. Robot technologies and applications are covered in chapter 3. Each of the chapters begins with a detailed roadmap of the relationships among the major subfields and topics featured in the papers for that chapter.

Cognitive intelligence is the process of selecting a particular behavior in a particular situation. The papers in chapter 1 discuss topics such as learning with neural networks, knowledge representation, and memory design, targeting applications such as obstacle detection, visual object categorization, intelligent operating system architectures, and prediction of user intentions. Cognitive intelligence is related to behavioral intelligence, which is defined as the hardware implementation of a desired behavior. Social intelligence considers topics such as human-robot interaction, robot-robot interaction, and robot collaboration in society. In this section, applications include the manifestation of facial expressions by a robot head, gesture exhibition, gaze control, remote operation, and human-robot cooperation in carrying an object. Note that social intelligence depends on behavioral intelligence because, in social intelligence, hardware specifications are carefully examined to make the robot safe in its interactions.

An interesting paper on behavioral intelligence is “Automatic Take-Off and Landing Control for Small Unmanned Helicopter” (pages 155 to 167), in which the author uses ultrasonic sensors to stabilize the altitude of a helicopter when it is near the ground, and an acceleration feedback control system based on a Kalman filter (linear quadratic estimation) to estimate horizontal acceleration. This control system for small unmanned helicopters achieves quite effective results in automatic take-off and landing. As a result, the author’s future work will consider fully autonomous flight for such helicopters.

The paper “Comparative Analysis of Arm Control Performance using Computational Intelligence” (pages 187 to 197) is also recommended. The authors studied different models of artificial neural networks to solve the inverse kinematics problem. The end part of a robotic arm, which interacts with the environment, is called the end effector, and a motion plan is required for the end effector to achieve its desired task. Different motion plan configurations exist for different navigations. The inverse kinematics problem asks the following question: What parameters (such as the angles of the joints) determine a specific desired configuration or position of the end effector? The authors compared feed-forward neural networks, neuro-fuzzy systems, and echo state networks (ESNs) to identify a solution to this problem. They determined that ESNs provide the best results.

In “Experience with the Children-Humanoid Interaction in Rehabilitation Therapy for Spinal Disorders” (pages 347 to 357), the Nao humanoid robot takes the role of a physiotherapist and shows successful results in rehabilitating children with scoliosis. In “Walking Pattern Generation on Inclined and Uneven Terrains for Humanoid Robots” (pages 209 to 221), the authors introduce a novel method based on modifiable walking pattern generation for stable walking and prove it to be effective. In “Humanoid Interface for Artificially Intelligent Role-Based Game Playing” (pages 283 to 291), humanoid robots called dynamic anthropomorphic robots with intelligence--open platform (DARwIn-OPs) play tic-tac-toe competitively with each other or with humans.

Ambient intelligence refers to the ability of a robot to gather data based on information provided by sensors that read the robot’s self-environment. Similarly, collective intelligence refers to the ability of the robot to acquire information from other robots that are operating cooperatively in its environment. These robots are considered as a source of sensory information, too. Genetic intelligence, which refers to robots that achieve computational processes with evolutionary algorithms, involves building a unique personality for a robot. The process is tailored to execute an anticipated complex behavior, which helps the robot achieve a desired objective. A robot’s knowledge can be represented with chromosomes based on genetic codes, which use soft computing to implement genetic algorithms to achieve their desired functional behavior. Such robots are regarded to have a form of genetic intelligence. Moreover, embodied and developmental robotics is the engineering paradigm to achieve such behavior. This is a fresh area for scientific research, which opens the door for novel scientific approaches and future prospects in robot intelligence.

Several interesting papers are found in chapter 2. In ambient intelligence, I recommend “Experimental Tests of Autonomous Jellyfish Removal Robot System JEROS” (pages 395 to 403), which uses a camera in an autonomous navigation system with a grid to detect, follow, and cut off intrusive jellyfish. In “Hybrid Indoor Location Tracking for Pedestrian Using a Smartphone” (pages 431 to 440), the authors use a localization method based on Wi-Fi (basically, Wi-Fi fingerprinting, which works by detecting the strength of a wireless signal) and sensors (specifically, inertial accelerometer sensors built into today’s smartphones). This is useful because global positioning systems (GPS) do not work indoors. Another paper, “Novel Scheme of Real-Time Direction Finding and Tracking of Multiple Speakers by Robot-Embedded Microphone Array” (pages 453 to 462), discusses how a robot can identify its master’s voice and differentiate its master’s voice commands from other speakers. In collective intelligence, the paper “Market-based Multiagent Framework for Balanced Task Allocation” (pages 549 to 559) discusses task allocation by assigning tasks based on the energy level of each robot. In genetic intelligence, “Distributed Multiobjective Quantum-Inspired Evolutionary Algorithm (DMQEA)” (pages 663 to 670) interrelates quantum computing, evolutionary computing, and parallel processing. In the embodied and developmental robotics section, highlights include “Personalized Emotional Expressions to Improve Natural Human-Humanoid Interaction” (pages 691 to 702), in which the authors use fuzzy logic to map user emotions onto the humanoid robot Nao, and “Computational Intelligence for Creating Autonomous Robots” (pages 733 to 740), which discusses theoretical concepts such as self-organization and self-configuration in the design of autonomous robots that can customize their abilities according to different users and different environments.

Chapter 3 tackles modern intelligent robot technologies and applications, such as robots playing soccer, robots in education, industrial robots, robots that can avoid collisions, and robots using ontologies as a way to communicate.

In summary, this book contains many technical and challenging papers, with an abundance of scientific material. It is impossible to mention all of them in this review. The chapters discuss diverse subjects and applications in computer software and simulation. Most of the papers contain math and physics, so a background in these fields is required. This book is a good reference for researchers involved in robotics and robot intelligence technologies, and anyone else with an interest in this area.