Action Prediction in Smart Home Based on Reinforcement Learning

Hassan, Marwa Yousif; Atieh, Mirna

doi:10.1007/978-3-319-14424-5_22

Cited by 6 publications

(6 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One of the most famous of these works is The Neural Network House [42] where a smart home was controlled by the ACHE system [43], which commanded the lights in order to reduce the overall energy consumption without impairing the inhabitant comfort based on the perceived environment. Although this project showed promising results and gave rise to other research [44], [45], the RL technique employed poorly scales to large real cases sets. As demonstrated in [9], a way to solve this issue, is to change the data structure, relying for example on a deep neural network [5], [6].…”

Section: Related Workmentioning

confidence: 95%

Context Feature Learning through Deep Learning for Adaptive Context-Aware Decision Making in the Home

Brenon¹,

Portet²,

Vacher³

2018

2018 14th International Conference on Intelligent Environments (IE)

View full text Add to dashboard Cite

In Intelligent Environments, prediction and decision must take the context of interaction into account to adapt themselves to the evolving environment. If most of the approaches to deal with this problem have used a formal representation of context, we present in this paper a direct extraction of the context from raw sensor data using deep neural network and reinforcement learning. Experiments undertaken in a voice controlled smart home showed which elements are useful to perform context-aware decision-making in the home and the adequacy of reinforcement learning to tackle an evolving environment.

show abstract

Section: Related Workmentioning

confidence: 95%

Context Feature Learning through Deep Learning for Adaptive Context-Aware Decision Making in the Home

Brenon¹,

Portet²,

Vacher³

2018

2018 14th International Conference on Intelligent Environments (IE)

View full text Add to dashboard Cite

show abstract

“…They employed the concept of activity probability and reward in reinforcement learning to infer if a detected activity is erroneous or not. Also, (Hassan & Atieh, ) proposed a reinforcement learning‐based action prediction in smart home that learns the change in user's behavior using the human action on devices as feedback.…”

Section: Background and Related Workmentioning

confidence: 99%

Situation‐centered goal reinforcement of activities of daily living in smart home environments

2019

View full text Add to dashboard Cite

Older adults with early‐stage dementia (ED) can experience confusion or lack clarity when performing routine activities of daily living (ADLs). These circumstances predispose the older adult to safety‐critical and often risky situations. A safety‐critical risky situation is one that constitutes a hazard. To support independent living, a sensor‐laden smart environment can be employed to mitigate such hazards. In this paper, we propose a situation‐centered goal reinforcement framework that supports older adults with ED in their decision making, and guides them through their ADL in order to fulfill their goal or intention and avoid hazards. First, we employ an LSTM (Long Short‐Term Memory) model to infer the current goal of the resident, using their previously observed normal ADL patterns. Secondly, we identify potentially risky situations in their currently observed goal path. We then incorporate a situ‐learning agent (SLA) that helps an inhabitant to make the right decision, thus preventing adverse events while guiding her through the task sequence that leads to her goal state. In addition, we use a naïve agent to simulate episodes of confusion similar to those that might be experienced by older adults with ED. We validated our method against an open‐source dementia dataset (Quesada et al., 2015) by considering four types of ADLs as case studies. We achieved an accuracy of 90.1% for our goal inference model, higher than the accuracies reported by related studies. We also reported other metrics including precision, recall and f1‐score for goal inference model. Finally, SLA's action recommendations relevance was evaluated accordingly.

show abstract

“…Other studies such as [33][34][35] do not use reinforcement learning. However, they use the Markov and clustering algorithms to learn which sets of activities and actions constitute a clear task and utilise them to predict the user actions in the smart environment.…”

Section: Related Workmentioning

confidence: 99%

“…The experimental results show the system outperforms other algorithms and gives more efficiency results which reduce memory usage and time. In contrast, Hassan [35] included a machine learning methodology and context-aware adjustment approach. At the high state, a schedule causes and learning framework is utilised to observe client and predict the future activities.…”

Section: Related Workmentioning

confidence: 99%

Hybrid user action prediction system for automated home using association rules and ontology

Shaban

Almasalha

Qutqut

2019

IET wirel. sens. syst.

View full text Add to dashboard Cite

Nowadays, with the rapid increase of Internet users, the Internet services dominate a primary part of our lifestyle. Moreover, the evolution of the internet of things has introduced new insights into smart platforms and devices that leads to the new vision of 'smart homes'. The idea of smart homes is not a recent concept; it has been in high interest by both academia and industry to make smart homes a more convenient technology for human's comfort. In this study, the authors propose a new hybrid prediction system based on the frequent pattern (FP)-growth and ontology graphs for home automation systems. Their proposed system simulates the human prediction actions by adding common sense data by utilizing the advantages of the ontology graph and the FP-growth to find a better solution in predicting home user actions for automated systems. For the evaluation of the proposed system, two ontology graphs are introduced with FP-growth to achieve the best results. Both graphs are tested through multiple weight values with the results of FP-growth. As a result, the best weight distribution selected in this study is (70, 30) for time and location ontology graphs respectively. Their results showed that the proposed prediction system achieved an accuracy of 79% for all weekdays and 81% excluding weekend days.

show abstract

Action Prediction in Smart Home Based on Reinforcement Learning

Cited by 6 publications

References 11 publications

Context Feature Learning through Deep Learning for Adaptive Context-Aware Decision Making in the Home

Context Feature Learning through Deep Learning for Adaptive Context-Aware Decision Making in the Home

Situation‐centered goal reinforcement of activities of daily living in smart home environments

Hybrid user action prediction system for automated home using association rules and ontology

Contact Info

Product

Resources

About