Uploaded byAhmad Lotfi
From Non-Intelligent to Intelligent Environments: a Computational and Ambient Intelligence Approach
The document discusses the transition from non-intelligent to intelligent environments, focusing on computational and ambient intelligence. It defines key concepts including artificial intelligence, machine intelligence, and ambient assisted living, while highlighting the potential of intelligent environments to support aging populations through innovative technologies. The document emphasizes the importance of monitoring and aiding elderly individuals, as well as creating user-friendly intelligent spaces that enhance daily living.
More Related Content
Similar to From Non-Intelligent to Intelligent Environments: a Computational and Ambient Intelligence Approach
![UiPath Automation Suite Installation (Hands-On) [2/3]](https://cdn.slidesharecdn.com/ss_thumbnails/automationsuitecommunitysession2-251015095633-a6d862f1-thumbnail.jpg?width=600ounds&width=560&fit=bounds)
From Non-Intelligent to Intelligent Environments: a Computational and Ambient Intelligence Approach
- 1. From Non-Intelligent toIntelligent Environments: a Computational and Ambient Intelligence Approach Inaugural Lecture by AHMAD LOTFI School of Science and Technology Nottingham Trent University
- 2. Spot the Difference! Non-IntelligentIntelligent 2
- 3. Spot the Difference#2 House BHouse A 3
- 4. Thermogram Spot the Difference#2 Less Intelligent More Intelligent
- 5. What is Intelligence? •Thecapacity to learn and solve problems, in particular: • the ability to solve novel problems •the ability to act rationally •the ability to act like humans 5
- 6. Test of Intelligence H0W0UR M1ND5 C4N D0 4M4Z1NG 7H1NG5! 1MPR3551V3 7H1NG5! 1N 7H3 B3G1NN1NG 17 WA5 H4RD BU7 N0W, 0N 7H15 LIN3 Y0UR M1ND 1S R34D1NG 17 4U70M471C4LLY W17H0U7 3V3N 7H1NK1NG 4B0U7 17. B3 PROUD! 0NLY C3R741N P30PL3 C4N R3AD 7H15. C4N YOUR COMP73R R34D 4ND UND3R574ND 7H15? 6
- 7.
- 8.
- 9. Story of theDoor Mechanical Intelligence Computational Intelligence 1 2 3 4 5 Artificial Intelligence Machine Intelligence 9
- 10. Machine Intelligence • ArtificialIntelligence (AI) - The study of computer systems that attempt to model and apply the intelligence of the human mind. • Computational Intelligence (CI) - Use of soft computing techniques to mimic the ability of human mind in effectively employing modes of reasoning that are approximate rather than exact. 10
- 11. Machine Intelligence Autonomy (self-diagnostics, faulttolerance, self-tuning) Man-Machine Integration (Human-like understanding/communication, emergence of emotion) Bio-inspired Behaviour (Cognitive-based, biologically motivated behaviour) Controllability (Adaptation, thinking and planning) System 1 System 2 11
- 12. Spot the Difference#3 Less Intelligent (IQ ≈ 100) More Intelligent (IQ ≈ 190) GeorgeWBush GarryKasparov 12
- 13. IQ Intelligence Quotient (IQ)is a composite indicator meant to measure people cognitive abilities in relation to their age group. 13
- 14. MIQ Machine Intelligence Quotient (MIQ)is a comparative indicator of machine human-like intelligence. 14
- 15.
- 16.
- 17.
- 18. 1. Reactive System SenseAct Smoke Alarm PIR Thermostat Light Heating/Cooling Alarm 18
- 19. Light Heating/Cooling Alarm 2. Deliberative System Think ActSense NEST– Learning Thermostat Thermostatic Radiator Valve 19
- 20.
- 21. Intelligent Offices Light Control Room Temperature PC/LaptopControl Office Occupancy 21
- 22. Definition • Intelligent Environmentsare spaces with embedded systems and information and communication technologies creating interactive spaces that bring computation into the physical world. • Ambient Intelligence refers to a digital environment that proactively supports people in their daily lives. 22
- 23. … or ...IntelligentEnvironments Ambient Intelligence AmbientAssisted Living Ubiquitous Computing Pervasive Computing Context Aware Internet of Things … Intelligent Offices Intelligent Cars Intelligent Cities … 23
- 24.
- 25. Definition • Ambient AssistedLiving (AAL) is the use of information and communication technologies in a person's daily living and working environment to enable individuals to stay active longer and live independently into old age. • This could be as simple as an alarm to remind a person to take medication or as sophisticates as a mobility scooter or electric wheelchair to help with daily shopping. 25
- 27. Sources: CORDIS, TSO“Carers of Elderly People – Summary of the Background Evidence” Alzheimer’s Society What is the Problem? • By 2050, people aged 65-79 are expected to make up almost 1/3 of the population in Europe • Over that period, the population of very elderly (80+) will rise by 180% • One in 50 people aged 65-70 have a form of dementia, rising to one in five over 80 • Currently there are 7 million carers of the elderly in the UK alone. 27
- 28. Tools & Infrastructure Whatmakes Ambient Assisted Living (AAL) possible? A. Assistive & Social Robotics • … a robot that interacts and communicates with humans or other devices … B. Wearable & Mobile Devices C. Ambient Assisted Living Homes • ... digital environment that proactively supports people in their daily lives 28
- 29. A. ASSISTIVE &SOCIAL ROBOTICS Tools & Infrastructure
- 30. Social Robots • Semi-autonomousrobots (mobile platforms, humanoid, …) • Interacting with people in their own space • Augmenting healthcare givers • Providing rehabilitation and lifestyle support 30
- 31. Social Robots (Examples) VGoAnybots QB DoubleRobot MantaroBot 31
- 32. Assistive Robots (Examples) UserAssisted Mobility Scooter Assisted RoboticsActive Walker http://iwalkactive.com 32
- 33. B. WEARABLE &MOBILE DEVICES Tools & Infrastructure
- 34. • Applications • Healthmonitoring • Navigation and stray prevention • Mobile persuasive technologies Wearable & Mobile Sensors iHealth - Wireless Activity and Sleep Tracker GPS Smart Shoe Smart CanePanic Alarm 34
- 35. C. AMBIENT ASSISTEDLIVING HOMES Tools & Infrastructure
- 36. Unobtrusive “Activities ofDaily Living” Monitoring System • To gather data on the routine activities of elders e.g., getting out of bed, going to the bathroom, preparing meals, taking medications etc. without altering the elders' normal behaviour • Preferably wireless sensors should be used, along with a small computerised receiver to collect data that are then analysed and posted to a secure central web site for viewing by the carer/relative • The adult children of frail elders living alone and at a distance can be sent reports or alerts daily/weekly in the form of e-mail or phone calls. 36
- 37. Ambient Assisted Living Homes Activities ofDaily Living (ADL) Kettle Microwave Fridge Doors Food cupboards Cutlery drawer Bed Room occupancy 37
- 38.
- 39.
- 40. Sensors … • FallDetector • Bed Occupancy Sensor • Chair Occupancy Sensor • Carbon Monoxide Alarm • Door Usage Sensor • Electrical Usage Sensor • Enuresis Sensor • Flood Detector • Gas Shut Off Valve Solution • Natural Gas Detector • Heat Detector • Medication Dispenser • Passive Infra-Red Movement Detectors • Pressure Mat • Pull Cord • Smoke Detector • Temperature Extremes Sensor 40
- 41. Ambient Assisted LivingHomes - the possibilities 41
- 42. Activities of DailyLiving (Example) http://www.tunstall-life.com/ 42
- 43. Occupancy Signal –Single Occupancy Time LoungeSensor Time KitchenSensor 0 1 • No parallel activity of PIRs • No uncertainty on where the occupant is. 43
- 44. Start-time and Duration BedroomKitchen Lounge Corridor 44
- 45. Start-time and Durationfor Bedroom Signal 45
- 46. Multiple Occupancy Scenario Noparallel activities Parallel activities 46
- 47. Hour Behaviour Pattern Room transition PatternRecognition Hour 47
- 48. Telehealth Solutions Remote exchangeof data between a patient at home and their clinician(s) to assist in diagnosis and monitoring typically used to support patients with Long Term Conditions. 48
- 49.
- 50. ACTIVITIES CLASSIFICATION AND ABNORMALITYDETECTION Research Area
- 51.
- 52. User Activities Outlier Detection •To detect any deviation in the day-to-day behavioural patterns of occupants using data generated from low level sensors. • To develop a good understanding of the normal behaviour and distinguish any abnormalities and possible trend in the behavioural changes. • To examine the application of distance measures and Fuzzy rule-based system in identifying the abnormality within the behavioural patterns of an occupant. 52
- 53. Case Studies • Case1 • This environment is monitored using JustChecking Monitoring system. • The data is collected for 14 months. • It is based on a single elderly occupant • Front and back door sensors, lounge, kitchen, bedroom, bathroom and upstairs motion sensors are used. • Case 2 • An elderly person was first prescribed some medications • After a few days, her medications were replaced • This environment is monitored using JustChecking Monitoring system. • Case 3 • The elderly person uses a walker support to help her in moving around her apartment. • Four motion sensors are used: lounge, kitchen, bedroom and corridor sensors. Two door entries are used: bathroom and the main entrance. • The data is collected for a couple of weeks where holidays and weekends are not included. 53
- 54. Extremely Outlier NormalMoreor Less Normal Scattered plot for the 1st and 2nd principal components of the back door entry sensor data used in case study I with classification. Results on Case Study (1) 54
- 55. Extremely Outlier NormalMoreor Less NormalSlightly Outlier Results on Case Study (1) Scattered plot for the 1st and 2nd principal components of the lounge motion sensor data used in case study I with classification. 55
- 56. Slightly OutlierNormal Moreor Less Normal Results on Case Study (2) Scattered plot for the 1st and 2nd principal components of the bedroom motion sensor data used in case study II with classification. 56
- 57. Extremely OutlierNormal Results onCase Study (3) Scattered plot for the 1st and 2nd principal components of the front door entry sensor data used in case study III with classification. 57
- 58.
- 59. Behaviour Modelling Techniques • HiddenMarkov Model (HMM) - is a statistical model in which the occupant behaviour is assumed to be a Markov process. • Recurrent Neural Network (RNN) - The neural network based approaches use large time series data sets to learn the relationship between the input data and output data. • The accuracy of both statistical and neural network based methods degrade rapidly with increasing prediction lead time. 59
- 60. 6 Hours AheadPrediction Using ESN Predicted values for bedroom occupancy sensor Predicted values for corridor occupancy sensor 60
- 61. Research Projects … •School Transport Automatic Register • Energy Efficiency in Social Housing • Activities Recognition and Worker Profiling in the Intelligent Office Environment • Intelligent Care Guidance and Learning Services Platform for Informal Carers of the Elderly (iCarer)
- 62. Problem Definition • Wewant the office environment to provide better worker comfort combined with reduced energy use • Setting the correct heating level • Setting satisfactory lighting • Only turning off computers when required and vice versa • We need to be able to identify the patterns of individuals’ Activities of Daily Work (ADW), so as to tailor their work environment to their particular needs. • Thus we have to investigate the similarities in behavioural patterns of different users in an office environment.
- 63. System Architecture • Intelligentoffice monitoring used 11 non-intrusive sensors. • Door and Window entry sensors • PIR • Temperature sensor • Humidity sensor • Chair pressure pad (vibration sensor) • Mouse and Keyboard activities System architecture for intelligent office environment System Architecture and Data Collection System
- 64. Can We Seea Pattern? • Aggregated data can appear complex
- 65.
- 66. Are the States‘Sensible’? • The office users were asked to generate a diary, and the contents compared with the basic states deduced from the sensors • This is promising, since it means that an Intelligent Office is unlikely to misinterpret states and annoy the office worker Tuesday, worker #2, diary of activities Tuesday, worker #2, sensor based states
- 67. Experiments / Results •Average/Moving Average are used to define the duration membership functions. • The worker clearly has different patterns on different days of the week • The coarse profile, considers only the average in order to provide the first stage of modelling • There will be some seasonal changes so that the pattern of work may very during the summer and winter time Sample of chair occupancy duration over three weeks.
- 68. Fuzzy Characteristics Matrix ChairActivities on Wednesday Chair Activities on Friday • Size of the circle represents the likelihood of the activity occurrence for the specified start time and duration for that day of the week.
- 69. Fuzzy Characteristics Matrix ChairActivities on Wednesday Chair Activities on Friday • Less important activity occurrence (small circles) could be discarded.
- 70. Research Projects … •School Transport Automatic Register • Energy Efficiency in Social Housing • Activities Recognition and Worker Profiling in the Intelligent Office Environment • Intelligent Care Guidance and Learning Services Platform for Informal Carers of the Elderly (iCarer)
- 71. Intelligent Care • Theinformal carers are becoming crucial agents in the elderly’s care and support. • Carers suffer distress or over-work episodes appeared due to lack of knowledge about older adult’s care. • Support and distress relief in caregivers should be a key issue in the home-care process of these older adults. • The project aimed at developing a personalized and adaptive platform to offer informal carers support by means of monitoring their activities of daily care and psychological state, as well as providing an orientation to help them improve the care provided. http://icarer-project.eu/ 71
- 72.
- 73. Concluding Remarks • AmbientIntelligence is foreseen to be present everywhere in the future world and to ease human living. • To build an environment which will be natural, informative and caring from human perspective. 73
- 74.
- 75.
- 76. T H AN K S