Sarmad Idrees

Learning humanoid skills from videos typically requires a successful human demonstration, which often demands custom data collection. Although failures have traditionally been treated only as negative examples in robot learning, they can still reveal a usable trajectory prefix before the task fails, as well as the intended outcome. To leverage this information from a failed-attempt video, we propose TRACC, a pipeline that imitates the useful portion of the motion trajectory and then completes the task based on the inferred task outcome. The usable motion prefix serves as prior knowledge until the failure occurs, after which the task-completion reward guides the policy to learn the intended task goal without requiring a successful task trajectory. We evaluate our method on six in-the-wild failed human tasks from the Oops! dataset. Our experimental results demonstrate the effectiveness of the proposed approach for learning from failed attempts when no successful demonstration is available. Thus, these findings establish failed human videos as a viable source of supervision for humanoid skill learning.

Many physical tasks in human environments require collaboration, from assisting a partner to jointly manipulating an object. Yet, existing humanoid benchmarks largely focus on single-humanoid skills and lack evaluation of multi-humanoid collaboration under egocentric visual observations. We introduce CoHuB (Collaborative Multi-Humanoid Benchmark), a simulation benchmark for multi-humanoid collaboration under egocentric visual observations. CoHuB provides 10 tasks, eight with two humanoids and two with three humanoids, spanning diverse collaboration patterns. We also provide synchronized demonstrations collected through a multi-operator VR teleoperation pipeline, in which each operator controls one humanoid from its egocentric view. Experiments with representative visuomotor policies reveal substantial challenges across different forms of coordinated perception and control. CoHuB provides a foundation for developing and evaluating multi-humanoid collaboration policies.

Human motion prediction is a crucial capability for advanced robotic systems that interact with humans. In facilities with dynamic human-robot collaboration settings, robots must anticipate human movements to ensure safety, prevent collisions, and optimize cooperative tasks. Traditionally, motion forecasting is treated as a sequential modeling problem using historical pose data, but achieving long-term accuracy and physical realism remains challenging. We present Adversarial Motion Transformer (AdvMT), a novel approach that integrates a Transformer-based motion encoder with a temporal continuity discriminator to address these challenges. The Transformer captures rich spatio-temporal dependencies across human joints, while adversarial training with a continuity discriminator enforces smooth, natural motion trajectories that adhere to biomechanical constraints. Our training scheme includes a bone-length consistency term and adversarial loss to reduce common artifacts like pose freezing or unnatural transitions. In experiments on the Human3.6M motion dataset, AdvMT achieves state-of-the-art long-horizon prediction accuracy while also delivering robust short-term predictions. These improvements strengthen the prediction foundation for physical AI in manufacturing and human-robot collaboration, where anticipating human motion is a prerequisite for safe and efficient robot coordination.

Predicting 3D human motion is a complex task, owing to the unpredictable nature of human movements. The influx of deep learning innovations and the availability of extensive datasets have intensified research interest in this field. This survey provides an exhaustive review of human motion prediction algorithms and categorizes them according to their core architectures, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Graph Convolutional Networks (GCNs), Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), Transformers, and Equivariant Neural Networks (ENNs). Our key contribution is a systematic presentation of the latest prediction methodologies, classified into direct and geometry-aware modeling. We begin with the problem formulation of human motion prediction, explore assorted techniques, and discuss data representation, accompanied by a list of accessible datasets. We also identify and analyze the ongoing challenges and limitations of the current algorithms, offering insights into potential future developments in this domain.

The workability of fresh concrete is highly important in terms of construction quality and safety. Slump tests are required every 120 m³, yet automated monitoring for each concrete batch remains unavailable in the actual concrete batching plant. To mitigate this issue, we propose an automatic slump prediction method based on the VGG16 neural network by analyzing the video from the final discharge hopper of the batching plant. Additionally, Explainable AI (XAI) is adopted to evaluate and validate our automatic concrete quality inspection approach. Iteratively examining XAI outputs and applying necessary adjustments in data preprocessing helps to achieve better overall performance. The proposed video classification method performed by averaging over the image-level predictions can classify the concrete into four slump classes with an average precision of 85% and an average F1 score of 87%. This demonstrates the possibility of continuous quality evaluation for all concrete produced in the concrete batching plant.