Surgical robot task automation has recently attracted great attention due to its potential to benefit both surgeons and patients. Reinforcement learning (RL) based approaches have demonstrated promising ability to perform automated surgical manipulations on various tasks. To address the exploration challenge, expert demonstrations can be utilized to enhance the learning efficiency via imitation learning (IL) approaches. However, the successes of such methods normally rely on both states and action labels. Unfortunately, action labels can be hard to capture or their manual annotation is prohibitively expensive owing to the requirement for expert knowledge. Emulating expert behaviour using noisy or inaccurate labels poses significant risks, including unintended surgical errors that may result in patient discomfort or, in more severe cases, tissue damage. It therefore remains an appealing and open problem to leverage expert data composed of pure states into RL. In this work, we present an actor-critic RL framework, termed AC-SSIL, to overcome this challenge of improving learning process with state-only demonstrations collected by an unknown expert policy. It adopts a self-supervised IL method, dubbed SSIL, to effectively incorporate expert states into RL paradigms by retrieving from demonstrations the nearest neighbours of the query state and utilizing the bootstrapping of actor networks. It applies similarity-based regularization and improves its prediction capacity jointly with the actor network. We showcase through experiments on an open-source surgical simulation platform that our method delivers remarkable improvements over the RL baseline and exhibits comparable performance against action based IL methods, which implies the efficacy and potential of our method for expert demonstration-guided learning scenarios. Code will be made publicly available at https://github.com/Jingshuai-cqu/AC-SSIL.

The ability to accurately track pedestrians is valuable for various application designs. Although pedestrian tracking has been investigated extensively and owns a well-suited sensing platform, the proposed solutions are far from being mature yet. Pedestrian tracking contains step counting and stride estimation two components. Step counting already has commercial products, but the performance is still unreliable and less trustworthy in practice. Stride estimation even stays in the research stage without ready solutions released on the market. Such a non-negligible gap between the long-term research investigation and technique's actual usage exists due to a series of crucial applicability issues unsolved, including design's vulnerability to interfering activities, extracting purely body's movement from mixed sensor signals, and parameter training without user's intervention. In this paper, we deeply analyze human's gait cycles and obtain inspiring observations to address these issues. We incorporate our techniques into existing pedestrian tracking designs and implement a prototype, PTrack, on LG smartwatch. We find that PTrack effectively enhances the system applicability and achieves promising performance under very practical settings.

In this article, we propose a mobile system, Aware , which turns our wearable or mobile device into a ruler. It can estimate the size of objects that could be large in size and not directly touchable by the user. Such a design will enable a rich set of applications that count on the size information of surrounding environments/objects. Aware purely utilizes the motion sensors on the device for object size measures. It can also integrate with the crowdsourcing feature for both performance improvement and result sharing. We propose a series of key techniques to address three major challenges in the Aware design: (1) user’s angle of line-of-sights to the object is used in the size measure but motion sensors track only the angle of arm’s waving, (2) motion sensors are noisy that require novel and effective data processing techniques, otherwise the errors could easily overwhelm the final result, and (3) in the crowdsourcing mode, Aware needs to identify vicinal objects of similar sizes and effectively fuse the measured sizes that correspond to the same object. We consolidate the above designs and implement Aware on Android platforms. Extensive experiments with four users show that Aware can achieve accurate measurement performance for the objects of various sizes in both indoor and outdoor environments.

To protect user privacy, many smartphone systems adopt the permission-based mechanism in which a user can evaluate the risk of requests for private information from a mobile app before installing it. However, recent studies show that the permission based mechanism is vulnerable to application collusion attacks because two apps, which appear to be harmless individually, can establish a covert channel and use it to leak confidential information. Consequently, people have designed some covert channel detection schemes, by checking abnormal status of the phone. In this paper, we point out that existing covert channel detection schemes may fail to detect a new type of collusion attacks referred as user-behavior-based covert channels. We implement three covert channels on Android smartphones. Our work sets a new alarm for the security issue of using smartphones. We then study the countermeasures to this new type of covert channels. Instead of trying to directly detect the proposed new type of covert channels, we propose two mitigation solutions to reduce the effectiveness of such covert channels. The mitigation solutions are also valid to other existing sensor-based side channels and/or covert channels on the phone.

The ability to accurately track pedestrians is valuable for variant application designs. Although pedestrian tracking has been investigated excessively and owned a well-suited sensing platform, the proposed solutions are far from being mature yet. Pedestrian tracking contains step counting and stride estimation two components. Step counting already has commercial products, but the performance is still unreliable and less trustworthy in practice. Stride estimation even stays in the research stage without ready solutions released on the market. Such a non-negligible gap between long-term research investigation and technique's actual usage exists due to a series of crucial applicability issues unsolved, including design vulnerability to interfering activities, extracting purely body's movement from additive sensor signals, and parameter training without user's intervention. In this paper, we deeply analyze human's gait cycles and obtain inspiring observations to address these issues. We incorporate our techniques into existing pedestrian tracking designs and implement a prototype, PTrack, on LG smartwatch. We find PTrack effectively enhances the system applicability and achieves promising performance under very practical settings.

In this work, we design and develop Montage for real-time multi-user formation tracking and localization by off-the-shelf smartphones. Montage achieves submeter-level tracking accuracy by integrating temporal and spatial constraints from user movement vectorestimation and distance measuring. In Montage, we designed a suite of novel techniques to surmount a variety of challenges in real-time tracking, without infrastructure and fingerprints, and without any a priori user-specific (e.g., stride-length and phoneplacement) or site-specific (e.g., digitalized map) knowledge: (1) a coded audio tone to support multi-user tracking with minimal latency, in the presence of high noise, multi-path effect, and Doppler Shift, (2) an innovative stride-length and walking direction estimation method without a priori knowledge of user and site, and (3) a vector-based multi-user tracking scheme which connects successive localization snapshots to refine users' locations and generate continuous moving traces. We implemented, deployed, and evaluated Montage in both outdoor and indoor environment. Our experimental results (847 traces from 15 users) show that the stride-length estimated by Montage over all users has error within 9cm, and the moving-direction estimated by Montage is within 20 degrees. For real-time tracking, Montage provides meter-second-level formation tracking accuracy with off-the-shelf mobile phones.

To protect user privacy, many smartphone systems, such as Android and Windows Phone, adopt the permission-based mechanism in which a user can evaluate the request of private information by a mobile app before installing it. However, recent studies show that the permission-based mechanism is vulnerable to application colluding attacks because two apps, which appear to be harmless individually, can establish a covert channel and use it to leak confidential information. In general, existing known covert channels usually work in a way that one app can modify the status of a system component, while the other can read the status. Even though several covert channel detection schemes have been proposed recently to fight against this type of covert channels, we point out that such designed covert channel detection schemes are not sufficient. In this paper, we demonstrate the possibility of establishing novel covert channels that work in quite different ways, in which one app (e.g., a game) can be designed deliberately such that the user will be induced to voluntarily modify the status of a system component (e.g., a motion sensor), while the other app can read the status of the system component. To validate our design, we implement three covert channels on Android. Our experiments show that these channels can bypass existing detection schemes. Moreover, we also measure the achievable throughput, error rate, and energy consumption in devices. The results demonstrate that our covert channels can achieve a transmission with high accuracy and low energy consumption. Our work sets a new alarm for the security issue of using smartphones.

In this work we design and develop Montage for real-time multi-user formation tracking and localization by off-the-shelf smartphones. Montage achieves submeter-level tracking accuracy by integrating temporal and spatial constraints from user movement vector estimation and distance measuring. In Montage we designed a suite of novel techniques to surmount a variety of challenges in real-time tracking, without infrastructure and fingerprints, and without any a priori user-specific (e.g., stride-length and phone-placement) or site-specific (e.g., digitalized map) knowledge. We implemented, deployed and evaluated Montage in both outdoor and indoor environment. Our experimental results (847 traces from 15 users) show that the stride-length estimated by Montage over all users has error within 9cm, and the moving-direction estimated by Montage is within 20o. For realtime tracking, Montage provides meter-second-level formation tracking accuracy with off-the-shelf mobile phones.